Nuclear Structure From

A Simple Perspective

R. F. Casten

OXFORD UNIVERSITY PRESS

NUCLEAR STRUCTURE FROM
A SIMPLE PERSPECTIVE
 OXFORD STUDIES IN NUCLEAR PHYSICS
General Editor: P.E. Hodgson

1. J. McL. Emmerson: Symmetry principles in particle physics (1972)

2. J. M. Irvine: Heavy nuclei, superheavy nuclei, and neutron stars (1975)
3. I.S. Towner: A shell-model description of light nuclei (1977)
4. P. E. Hodgson: Nuclear heavy-ion reactions (1978)
5. R. D. Lawson: Theory of the nuclear shell model (1980)
6. W. E. Frahn: Diffractive processes in nuclear physics (1985)
7. S. S.M.Wong: Nuclear statistical spectroscopy (1986)
8. N. Ullah: Matrix ensembles in the many-nudeon problem (1987)
9. A. N. Antonov, P. E. Hodgson, and I. Zh. Petkov: Nucleon momentum
and density distributions in nuclei (1988)
10. D. Bonatsos: Interacting boson models of nuclear structure (1988)
11. H. Ejiri and M. J. A. de Voigt: Gamma-ray and electron spectroscopy in
nuclear physics (1989)
12. B. Castel and I. S. Towner: Modern theories of nuclear moments (1990)
13. R. F. Casten: Nuclear structure from a simple perspective (1990)
NUCLEAR STRUCTURE FROM
A SIMPLE PERSPECTIVE

R. F. Casten
Brookhaven National Laboratory
Institutfur Kernphysik, University ofKoln

New York Oxford

OXFORD UNIVERSITY PRESS
1990
 Oxford University Press
Oxford New York Toronto
Delhi Bombay Calcutta Madras Karachi
Petaling Jaya Singapore Hong Kong Tokyo
Nairobi Dar es Salaam Cape Town
Melbourne Auckland
and associated companies in
Berlin Ibadan

Copyright © 1990 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.,
200 Madison Avenue, New York, New York 10016
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Casten, R.
Nuclear structure from a simple perspective/R.F. Casten/
p. cm.—(Oxford studies in nuclear physics)
ISBN 0-19-504599-8
1. Nuclear structure. 2. Nuclear excitation.
3. Collective excitations. I. Title. II. Series.
QC793.3.S8C37 1990
539.V4—dc20 89-23064 CIP

246897531
Printed in the United States of America
on acid-free paper
 PREFACE

This is a very personal book, reflecting the way I see and understand nuclear
physics and my belief that if something cannot be explained simply it is not
really understood, nor will it be a fertile inspiration for new ideas. It is quite
different from existing texts. It is in no way intended to replace them, but aims
instead to complement the standard approaches to nuclear structure.
The idea for the book and much of its approach arose gradually out of both
formal and informal courses in nuclear structure I have given at the Institute
Laue-Langevin in Grenoble, France, the Institut fur Kernphysik in Koln, Ger-
many, and at Drexel University, Philadelphia, as well as from a series of very
informal tutorial-like sessions with several of my graduate students. The book
represents an attempt to cut through the often heavy mathematical formalism
of nuclear structure and to present the underlying physics of some pivotal
models in a simple way that frequently emphasizes semi-classical pictures of
nuclear and nucleonic motion and repeatedly exploits a few fundamental
ideas. Such an approach has worked for me. I can only hope others will find
it useful.
The emphasis in this book always centers on seeking the essence of the
physics: rigor is therefore often sacrificed. Of course, rigor is absolutely
essential to a proper development of nuclear models and to precise calcula-
tions. Yet it can also be terrifying with page upon page of complex formalism,
Racah algebra, tensor expansions, and the like. Necessary as this is, especially
to those who will become practitioners of particular models, many readers can
become either discouraged or buried in the formalism. Unfortunately, in
either case, the beauty, elegance, and conceptual economy of nuclear structure
theory is often missed. The complexity necessary in formal treatments at times
obscures rather than illuminates the simple physics at work. Moreover, rigor
can be found elsewhere, in many excellent texts: there is no need for another
book to repeat it. (Several of the best of these texts, such as those by de Shalit
and Talmi, Bohr and Mottelson, Brussaard and Glaudemans, deShalit and
Feshbach, or the recent work by Heyde, are cited in the reference list at the
end. They are indispensable.) What cannot be found so easily, though most in-
dividual parts of it probably exist scattered throughout the literature, is a
systematic attempt to convey a more physically intuitive way of thinking about
nuclear structure and of extracting the essential physics behind the derivations
and models.
vi Preface

I honestly do not know whether this attempt will work. I feel it is worth a try
since it is, in fact, a way of thinking used every day by practicing scientists but
that is seldom presented in formal texts. If successful, it can deepen real
understanding. As T. D. Lee once said (BNL Colloquium, May 1983): "That a
thing is elementary does not mean it cannot be deep." Moreover, since physics
research is fundamentally a creative act depending on insight and imagination
(backed, of course, by hard science—experimental or theoretical) an intuitive
sense of nuclear structure can foster new inspirations and remove much of the
mystery surrounding formal or calculational complexity. Finally, an approach
of this type can be of considerable practical use. My hope in fact is that the
reader will come to appreciate how far one can go in obtaining many results of
detailed calculations almost instantly, essentially "by inspection." It can help,
for example, in anticipating the potential usefulness of a given model, in
spotting errors in calculations, or in estimating the effects of particular pa-
rameter changes. One of the best examples is the famous Nilsson diagram:
nearly all its features, and even the semiquantitative structure of Nilsson wave
functions, can be deduced without calculation. The same applies to much of
the study of residual interactions in the shell model, to collective models, to the
structure of RPA vibrational wave functions, Coriolis coupling, or the IB A. To
facilitate the development of this "sense" of the physics, and to provide contact
between models and real data, concrete examples are almost always discussed
in some detail and compared to exact calculations.
The book is in no way intended as a complete treatise on nuclear structure,
either in overall coverage or within each topic. Other texts are more compre-
hensive. For example, many active areas of modern nuclear physics (e.g.,
relativistic heavy ion physics and quark-gluon plasmas, mesonic and quark
degrees of freedom, or baryon excitations such as delta resonances) are totally
ignored. Nevertheless, the book relates to all of them in that it discusses the
basic models of nuclear structure, which successfully describe virtually all low
energy nuclear phenomena, and which subnucleonic approaches must eventu-
ally reproduce. Just as it is difficult to discuss relativistic effects in nuclei
without first knowing how far a non-relativistic theory can go, it is necessary to
understand how far traditional nuclear structure theory can go if one is to
isolate the effects of quarks, mesons, nucleonic excitations, and the like.
Even within traditional nuclear structure, many areas are bypassed. One
reason for this selectivity is that it allows deeper and more detailed discussions
of the subjects that are treated, discussions that would normally be found only
in specialized monographs. A more practical and honest reason is that the
book reflects that small part of the subject with which I am somewhat familiar
and where I felt I could attempt something that was more than just parroting
existing texts.
Finally, in addition to selectivity in topics there is a selectivity in treatment
of each subject. For many areas, there exist several equivalent approaches.
For example, there are works on residual interactions in the shell model that
barely mention the concept of seniority and others that stress it throughout.
Likewise, high spin phenomena in deformed nuclei can be discussed in terms
 Preface vii

of the Nilsson model with Coriolis coupling or with the formalism of the
cranked shell model. In each case, I have used the approach (in these
examples, seniority and the Nilsson-Coriolis concepts, respectively) that makes
the essential physics clearest to me (and, I hope, to the reader) and with which
I feel most comfortable.

Upton, New York R. C.

This page intentionally left blank
 ACKNOWLEDGMENTS

This book, while my responsibility, owes much to many people, both to those
who molded my own education or who were important in the development of
the approach I use here, and to those who played specific roles in the book
itself. I cannot mention all but would particularly like to single out, for special
thanks and gratitude, the following:

Above all, my father, Daniel E Casten, for everything.

Ed Kennedy, who gave me my initiation into the excitement of nuclear
physics.
D. A. Bromley, who more than fulfilled those early expectations and who
has continually encouraged me since.
D. D. Warner, who never lets me get away with anything, for his creativity
and insight, and for years of exciting work together at a perfect time. The
section of Chapter 6 on the IB A is only slightly modified from portions of a
review article co-authored with him. A particular piece of Chapter 7 (so
noted at the appropriate point) is also largely due to him.
F. lachello and A. Arima whose work and inspiration forever altered the
course of my own work.
Igal Talmi, for innumerable discussions and for making occasional friendly
disagreements so interesting, and enlightening.
J. A. Cizewski, who started the whole business of this book by asking and
caring deeply about nuclear structure and who seemed to enjoy my efforts
to answer.
A. Aprahamian, whose excitement at learning and discovery is an inspira-
tion, for our collaborations, friendship, and discussions.
Jolie and Ani were my principal "guinea pigs" in all this and I am very
grateful for their enthusiasm, which inspired my own, both for research and
for trying to understand what nuclear structure is all about. Without them,
this book would not exist.
K. Heyde, for many clarifications of my muddled thinking and collabora-
tions that taught me much. Part of Section 4.4 is based largely on discus-
sions with him, and Figures 4.14 a and b are based on ones in his book.
O. Hansen, S. Pittel, A. ("Aleee") Frank, D. S. Brenner, and Zhang, Jing-ye
for many productive and enlightening collaborations and discussions that
broadened my horizons.
x Acknowledgments

J. Weneser, M. Goldhaber, W. R. Kane, and H. Feshbach for discussions,

encouragement, and moral support.
H. Borner and K. Schreckenbach for a series of productive collaborations,
for demonstrating to me the power of precision, and for their efforts to
create such an attractive ambiance for nuclear physics at the ILL.
The students at the ILL in 1975-1976, those (particularly K. Schiffer) in
Koln in 1984-1985, and those at Drexel University in 1987 who also bore the
burden of listening to these ideas in embryonic form.
P. von Brentano and A. Gelberg for countless thought-provoking discus-
sions and for creating the stimulating atmosphere in Koln where much of
this book was written.
Gritti's Cafe, Dilrenerstrasse, Koln, and its staff, where most of that writing
actually took place, for its atmosphere, its music (by Achille), and its cap-
puccino conducive to physics and physics poetry.
The mountains near Wengen, Switzerland, for much of the same, at an
especially crucial stage.
My wife, Jo Ann, for her patience during thousands of hours when I was
trying to write this book and for submerging her feelings that nuclei are,
after all, a silly sort of thing to get all excited about.
Bob and Marian Haight and Don, Jo Ann and Kim Reisler, and Tom and
Ahng Suarez for years of friendship, talk, hiking, skiing, tennis, and general
support.
Jackie Hartt, whose patience is outdone only by her tennis.
Judy Otto, Irma Reilly, Walter Palais, and Whitey Caiazza of the Illustra-
tion group at Brookhaven National Laboratory for their very professional,
and enthusiastic, help with the figures. I am also grateful to the Photogra-
phy Group for their excellent work and for putting up with so many
"emergencies."
Jackie Mooney, whose skills, including cryptology, are absolutely amazing
and without whose work, friendship, and dedication this book would have
been sheer fantasy. I promise never to inflict such agony again.
I would also specifically like to thank K. Heyde, M. Buescher, W. Krips, R.
Schuhmann, and Zhang, Jing-ye for very careful reading of many parts of
the manuscript and for pointing out numerous errors.
I am very happy to acknowledge Brookhaven National Laboratory for
permission to pursue this project and for the infrastructure that made it
possible.
Finally, I am grateful to Oxford University Press for asking me to write this
book, for forbearance during several delays and lapsed deadlines, and for
making the whole process as pleasant as it has been.
 NOTATION

In a work such as this, the question of notation always presents a dilemma—

whether to adopt a rigorously unique set of symbols or to use those commonly
found in the literature. With one significant exception (so noted at the
appropriate place in the text), I have followed the latter course since it
facilitates further study by the reader and because, in practice, there should be
little ambiguity of meaning. Generally, different uses of a given symbol are
widely separated in context and location (with one awful exception in Chapter
4 concerning the "A" dependence of the interaction strength). To further help
avoid confusion it may be useful to list a few of the notational choices.
Y refers to a physical wave function while 0 usually refers to a basis state or
an unperturbed wave function.
E is used for excitation energies but also for quasiparticle energies.
e is used for single particle shell model energies and for one of the quad-
rupole deformation parameters.
Prefers to the "contact like" residual interaction and also to a quadrupole
deformation parameter.
Generally p, n are used for proton and neutron (although very occasionally
n and v are substituted). However, p also refers to particle as is p-p or p-h
(particle-particle or particle-hole excitation).
TV refers alternately to neutron number, boson number in the IB A, and to
oscillator shell number.
A is the mass number of a nucleus but also the angular part of a residual
interaction (especially in Chapters 4 and 5).
a is an amplitude of a mixed wave function and also the alignment quantum
number in the rotation alignment scheme.
J is used to denote the total angular momentum quantum number (colloqui-
ally, "the spin") of a level. Note that, quantum mechanically, the actual total
angular momentum is V77/+7) h. j is used for the angular momentum of a
single particle.
Finally, operators, matrices, and vectors are given in bold face. The same
symbol (e.g., Q, nj in normal face stands for the eigenvalue of that operator.
This page intentionally left blank
 CONTENTS

PARTI INTRODUCTION

1. INTRODUCTION, 3
1.1 Introduction, 3
1.2 The Nuclear Force, 6
1.3 Pauli Principle and Antisymmetrization, 16
1.4 Two-State Mixing, 17
1.5 Multistate Mixing, 22
1.6 Two-State Mixing and Transition Rates, 25

2. THE NUCLEAR LANDSCAPE, 28

PART II SHELL MODEL AND RESIDUAL INTERACTIONS

3. THE INDEPENDENT PARTICLE MODEL, 47

4. THE SHELL MODEL: TWO-PARTICLE CONFIGURATIONS, 67

4.1 Residual Interactions: The 5-Function, 69
4.2 Geometrical Interpretation, 85
4.3 Pairing Interaction, 93
4.4 Multipole Decomposition of Residual Interactions, 94
4.5 Some Other Results (Average Shifts, Hole, and Particle-
Hole Configurations), 104

5. MULTIPARTICLE CONFIGURATIONS, 109

5.1 J Values in Multiparticle Configurations: The m-Scheme, 109
5.2 Coefficients of Fractional Parentage (CFP), 111
5.3 Multiparticle Configurations;": The Seniority Scheme, 115
5.4 Some Examples, 125
5.5 Pairing Correlations, 130
xiv Contents

PART III COLLECTIVITY, PHASE TRANSITIONS,

DEFORMATION

6. COLLECTIVE EXCITATIONS IN EVEN-EVEN NUCLEI:

VIBRATIONAL AND ROTATIONAL MOTION, 141
6.1 An Introduction to Collectivity, Configuration Mixing,
and Deformation, 141
6.2 Collective Excitations in Spherical Even-Even Nuclei, 145
6.3 Deformed Nuclei: Shapes, 161
6.4 Rotations and Vibrations of Axially Symmetric
Deformed Nuclei, 165
6.5 Axially Asymmetric Nuclei, 186
6.6 The Interacting Boson Model, 197
6.7 The Development of Collectivity: Phenomenology and
Microscopic Basis, 234

7. THE DEFORMED SHELL MODEL OR NILSSON MODEL, 247

7.1 The Nilsson Model, 248
7.2 Examples, 260
7.3 Prolate and Oblate Shapes, 263
7.4 Interplay of Nilsson Structure and Rotational Motion, 264

8. NILSSON MODEL: APPLICATIONS AND REFINEMENTS, 271

8.1 Single Nucleon Transfer Reactions, 271
8.2 The Coriolis Interaction in Deformed Nuclei, 280
8.3 Coriolis Mixing and Single Nucleon Transfer Cross Sections, 290
8.4 Coriolis Effects at Higher Spins, 300

9. MICROSCOPIC TREATMENT OF COLLECTIVE

VIBRATIONS, 313
9.1 Structure of Collective Vibrations, 314
9.2 Examples: Vibrations in Deformed Nuclei, 322

PART IV EXPERIMENTAL TECHNIQUES

10. A FEW SELECTED EXPERIMENTAL TECHNIQUES, 335

10.1 Coulomb Excitation, 335
10.2 Spectroscopically Complete Techniques, 344
10.3 Heavy Ion Compound Reactions and High Spin States, 356
10.4 Heavy Ion Transfer Reactions, 361

REFERENCES, 365
INDEX, 371
 PART I

INTRODUCTION
This page intentionally left blank
 1
INTRODUCTION

1.1 Introduction
The atomic nucleus is not a single object but a collection of species ranging
from hydrogen to the actinides, and displaying an unbelievably rich and
fascinating variety of phenomena. The nucleus is extremely small, namely
about 10 ~12 to 10~13 cm in diameter, and can contain up to a couple of hundred
individual protons and neutrons that orbit relative to one another and interact
primarily via the nuclear and Coulomb forces. This system may seem so
complex that little could ever be learned of its detailed structure. Indeed,
many of us involved in research into nuclear structure proclaim loudly and
strenuously that we have barely scratched the surface (both literally and
figuratively, as we shall see) in our understanding of nuclear structure. From
another perspective, however, we have an immense number of facts about
nuclei and we understand an enormous amount, often in great detail, concern-
ing what the individual nucleons do in atomic nuclei, how this leads to the
observed nuclear phenomena, how and why these phenomena change from
nucleus to nucleus, and how certain nucleons interact with each other in the
nuclear medium. We have basic models—the shell model and collective
models—both geometric and algebraic, that provide a framework for our
understanding and that are extremely simple, and yet subtle and refined. It is
only after these models and framework are appreciated that the limitations in
our knowledge become focused and identifiable; the identification of the
problems that persist is a prerequisite to further advancement. In this book we
emphasize the known and understood as a background, map, and guide to the
unknown.
We hope the reader perusing this book will come to appreciate two prin-
ciple facts: namely, the beautiful richness and variety of nuclear physics and the
extent to which we can understand nuclear data and models by invoking a few
extremely basic ideas and drawing upon arguments that are physically trans-
parent and intuitive. We will see that it is possible, in many if not most cases,
to understand the detailed results of complex calculations with an absolute
minimum of formalism and often by inspection. As an example, even such
seemingly complicated results of nuclear models such as the famous Nilsson
diagram and the detailed structure of Nilsson wave functions, or of the
microscopic RPA wave functions of collective vibrations, can be derived,
nearly quantitatively, without any calculation whatsoever.
3
4 Introduction

We emphasize that one must make a careful distinction between this

approach and what is commonly called handwaving. The latter, to this author's
mind, is what one does when one does not really know an answer or explana-
tion and tries to explain some piece of nuclear data or the result of some
calculation by an offhand, qualitative, "explanation" that is often little more
than a slogan or a repetition of key words or venerated phrases. We have all
encountered examples of such handwaving: supposedly forbidden y-ray transi-
tions glibly explained as "due to mixing," extra or unexpected excited states
dismissed as "due to neglected degrees of freedom," unexplained peaks in
transfer reaction spectra ascribed to "higher order processes in the reaction
mechanism," or explanations of model calculations as "resulting from the
symmetry properties," or "from an energy minimization" (of course but why,
how?). Indeed, in many cases, such statements are true and are reasonably
accurate descriptions. Otherwise they would not have become catch phrases.
But abuses abound to such an extent that they often represent a watered down
substitute for real understanding that is to be discouraged.
The approach here, in contrast, attempts to extract the basic physics ideas
that emerge either from an inspection of nuclear data or that lie behind the
results of some model or calculation, and to do this by applying a minimum of
key physical and geometrical ideas about the nucleus. When attempting such
a program, there is always the danger of losing sight of important subtleties
and of ignoring the importance of formal rigor that is so necessary in detailed
and realistic model calculations. Undeniably, there are certain results of such
calculations and certain model predictions that can only be understood by
carrying out the fully detailed, often tedious, calculations. However, the
author has always felt, and hopes that the reader will be convinced, that it is
remarkable how far one can go in understanding complex nuclear structure
phenomena by careful but simple physical arguments. Of course, this ap-
proach has the lurking danger of itself slipping into handwaving. If such sins
are kept to the minimum here, there is a chance that the reader may emerge
with an appealing physical understanding and intuition into nuclear structure
that many working physicists have attained but that is seldom spelled out in
textbooks because of a hesitancy or reluctance to commit to writing the
nonrigorous and intuitive arguments that all of us use and, mentally, rely upon.
If anything, the philosophy of this book is that such ideas and such an approach
should not be a skeleton hidden in a closet for fear of ridicule but rather an
important aid that is a constant and continual complement to necessary formal
and rigorous calculations. As frequently as possible, qualitative physical
explanations or "derivations" will not be left to stand alone; rather, the
physical intuition behind them will be tested by, confirmed, and confronted
with the results of actual calculations or with the data on atomic nuclei
themselves.
We have already stated that many of the arguments here will rely on "a few
basic simple ideas." In fact there are three of these that are of absolute and
essential importance. They are:
• The generally attractive and short-range nature of the nuclear force.
 Introduction 5

• The effects of the Pauli principle.

• An understanding of two-state mixing—that is, the effects on energies
and wavefunctions when two nuclear states mix due to some residual
interaction.
These ideas, plus a constant reference to a kind of geometrical picture of the
nucleus, form the basic ingredients behind many of the arguments to be
presented.
A general outline of the book is as follows: After the discussion of these
three basic ingredients in Chapter 1, Chapter 2 departs from the usual way of
presenting nuclear structure by "surveying the nuclear landscape," that is, by
collecting a number of examples of nuclear data, level schemes, transition
rates, systematics, and so on. In a normal text that relies on a systematic, step-
by-step progression of ideas, such a chapter would be out of place since it
utilizes terminology and concepts that will be formally introduced later. It is
inserted here so that the beautiful and elegant consequences of nuclear models
discussed in subsequent chapters will not be presented in a contextual vacuum.
Often such results, when first encountered, seem highly abstract and of little
practical importance and it is, unfortunately, often only years later, when the
practicing nuclear physicist has gained a deeper familiarity with nuclear data,
that their significance is finally understood.
One reason we feel justified in this approach is that this book is aimed not
only at beginning nuclear physics graduate students, to whom much of the
jargon in Chapter 2 will be unfamiliar, but also to practicing and experienced
nuclear physicists who may be interested in the kind of alternative and comple-
mentary approach emphasized here.
Chapters 3,4, and 5 will deal with the shell structure of nuclei, as is tradi-
tional, starting with the independent particle model and going on to the shell
model for multiparticle configurations. The formalism and mathematical
development of the shell model is one of the most remarkable creations of
nuclear physics and allows one to account for many empirically observed
features of atomic nuclei with an absolute minimum of physical input (e.g.,
many detailed predictions can be made without ever specifying the nature of
the central shell model potential or the detailed structure of a residual interac-
tion). Unfortunately, the shell model formalism, and derivations based on it,
are often complex and, quite honestly, terrifying. This has the unfortunate
consequence that this subject is often skimmed or glossed over by students.
These chapters attempt to highlight and give plausibility arguments for many
shell model results while at the same time avoiding, as much as possible, such
formal treatments. Simpler derivations are sometimes possible and are given
in appropriate cases.
The next section of the book (Chapter 6) deals with collective models for
even-even nuclei, starting with macroscopic models of vibrational and rota-
tional motion. This material also emphasizes the profound importance of the
residual proton-neutron interaction (especially the T= 0 component) and its
role in the onset of collectivity, configuration mixing, and deformation in
nuclei, in inducing nuclear phase transitions, and the assistance its understand-
6 Introduction

ing can provide in simplifying the systematics of nuclear data. Following this,
a brief treatment of algebraic models, principally the interacting boson ap-
proximation (IB A) model is given.
In Chapters 7 and 8, the discussion turns to odd mass nuclei, with primary
emphasis on deformed nuclei, and an extensive discussion of the Nilsson
model and its consequences, extensions, and testing via single nucleon transfer
reactions.
Most of the collective models discussed up to this point will have been
phenomenological or macroscopic. We will not discuss detailed microscopic
approaches at length but will introduce such approaches in Chapter 9 since
they provide both the microscopic justification of macroscopic models and a
simple physical picture of collective excitations (especially vibrations) that will
allow the reader to anticipate their detailed structure without calculation.
Finally, in Chapter 10, a few selected experimental techniques for studying
atomic nuclei will be discussed, primarily to give the reader a glimpse of the
richness of experimental probes available, of their differences and of the
different types of nuclear data and structure that they elucidate, and to provide
some simple physical ideas that may be helpful in understanding the "parame-
ters" that govern the design of such experiments and the extraction of informa-
tion from them.
With this discussion of the philosophy and outline of the material to follow,
we turn now to the three "cornerstones" mentioned earlier that are of such
central importance to everything that follows. Many readers know that the
nuclear force is attractive, that the Pauli principle is important, and understand
that residual interactions can mix neighboring states. They might be tempted
to skip over these sections and of course that is their prerogative; indeed, these
pages contain nothing that is new or not to be found elsewhere. However, they
do present a somewhat different perspective and provide a ground and back-
ground for what follows.

1.2 The Nuclear Force

Nuclei exist and are composed of neutrons and positively charged protons.
Were the nuclear environment dominated by the repulsive Coulomb force,
this could not be the case: one therefore deduces immediately that there must
be a strong, attractive, interaction that can overcome the repulsive Coulomb
force and bind nucleons together. The nuclear force is at first glance a
mysterious one since it has few if any recognizable consequences in macro-
scopic matter (i.e., everyday phenomena). And, in fact, the exact nature of this
force is still largely unknown. Nevertheless, it is remarkable how much we can
learn about it from a few simple empirical facts. We have already stated that
the very existence of nuclei implies a new force—the strong interaction that
can overcome the Coulomb repulsion between protons. Beyond this many
experiments point to two basic facts:
• Nuclei are small, on the order of 10 ~12 to 10~13 cm in diameter.
 Introduction 7

Fig. 1.1. Binding energies per nucleon. The solid curve is the result of a typical semi-empirical
mass formula that includes corrections for surface effects, Coulomb repulsion, the Pauli principle,
and pairing effects. Many of these topics will be discussed later on. (Based on Eisberg, 1974.)

• For all practical purposes, the nuclear force can be neglected when
considering atomic and molecular phenomena.
These two facts tell us that the nuclear force must be short range. A few
further empirical observations allow us to refine this considerably.
• Nuclear binding energies, per nucleon, at first increase rapidly with A,
until about A ~ 10 to 20, where they level off at approximately 8.5 MeV
and remain roughly constant thereafter. These binding energies per
nucleon are shown in Figure 1.1.
• The masses of mirror nuclei, which are defined as pairs of nuclei with
interchanged numbers of protons and neutrons, (Z, N\ = (N, Z)2, are
nearly identical, after correcting for the different strengths of the Cou-
lomb interaction in the two nuclei.
• The sequencing, spin parity (/*) values, and excitation energies of excited
states in mirror nuclei are also nearly identical.
• Proton and neutron separation energies, denoted S(p) and S(n), are
defined as the energies required to remove the last proton or neutron to
infinity, and have a characteristic behavior with changing proton number
and neutron number. Typical examples of such separation energies are
8 Introduction

Fig. 1.2. Neutron separation energies near the N = 82 magic number (de Shalit, 1974).

shown in Figs. 1.2 and 1.3 from which it is evident that S(p) decreases
with increasing Z and increases with increasing N while S(n) decreases
with increasing N and increases with increasing Z. That is, each decreases
with an increasing number of the same type of nuclcon and increases with
an increasing number of the olherlype. Although not exactly germane to
the present discussion, we note for later use that the separation energies
also show particularly large and sudden drops at certain special numbers
of protons and neutrons, called magic numbers, namely N or Z = 2,8,20,
40, 50, 82,126. Those at 82 are evident in Figs. 1.2 and 1.3.
Measurements of electron scattering provide abundant evidence of a
nearly constant nuclear density independent of the number of nucleons
A. This, in turn, implies that the nuclear volume must increase linearly
with A. Neither of these facts may seem particularly surprising at first
but it should be recalled that such is not the case with atomic systems
whose sizes are nearly independent of Z. Note that if the nuclear volume
V °t A, then, assuming a roughly spherical nucleus, the nuclear radius
scales as Am. Innumerable studies have shown that a good approxima-
tion to the nuclear radius is R = R Aw where R -1.2 fin.
 Introduction 9

Fig. 1.3. Proton separation energies near the Z = 82 magic number (de Shalit, 1974).

• There is only one bound state of the deuteron, the simplest nuclear
system, with one proton and one neutron.
• This bound state has total angular momentum 7 = 1 , that is, in the
deuteron, the intrinsic spins (1/2) of the neutron and proton are aligned
parallel to each other, not antiparallel. (This result assumes that the two
nucleons have no relative orbital angular momentum, a result that will be
justified in Chapter 3.)
• The deuteron has a nonzero quadrupole moment, that is, it has, on
average, a preference for a nonspherical shape.
Let us consider each of these facts in turn and see what we can learn about
the nuclear force. Note that, except for the listing of the magic numbers, the
10 Introduction

preceding list contains virtually no specific nuclear structure information,

although what we will learn from these facts about the nuclear force has many
consequences for nuclear structure throughout the periodic table.
We have already deduced that the nuclear force is predominantly attractive
and short range. (We neglect the short range repulsive core component.) The
binding energy results, Fig. 1.1, and the fact that the nuclear density is approxi-
mately constant, allow us to go much further. The nuclear size, corresponding
to a radius of ~ IQ-12 to 10~13 cm, tells us that the range of the force must be
roughly this order of magnitude. However, the density and binding energy
data tell us that the force is actually of a considerably smaller range than that
of most nuclei. If the nuclear force extended more or less equally to all A
nucleons, then the binding energy would increase roughly as A (A -1)/2, or the
binding energy per nucleon (B.E.AA) would increase with A and therefore so
would the nuclear density. That such is not the case, at least for A > 10, shows
that the nuclear force saturates. The contrary fact that the binding energy per
nucleon does increase for very light nuclei allows us to quantify this saturation
and to make at least a crude estimate of the number of nucleons with which
each other nucleon interacts.
To do this, we assume that all such interactions are equal and count the
number of interactions assuming that each nucleon interacts with various
numbers of other nucleons. Fig. 1.4 gives a pictorial illustration of the connec-
tions and shows a plot of B.E./A deduced under different assumptions for the
numbers of connections per nucleon. If one works through this figure, it
becomes obvious how saturation arises when each nucleon interacts only with
a finite number, n., of others. For example, if «. = 3, the number of interactions
of each nucleon (and hence its binding energy) has already reached its maxi-
mum value when A = 4. We can use this approach to set some rough limits on
nr Clearly n. = 2 is unacceptable since it leads to immediate saturation at A = 3.
Similarly n. = 3 leads to premature saturation, but somewhere on the order of
6 to 10 mutual interactions leads to B.EJA values that approach saturation
roughly where the data do. Thus, the empirically observed saturation in
binding energy per nucleon data suggest that the range of the nuclear force is
on the order of the size of nuclei such as Li or Be (i.e., approximately 2 to 4 fm).
Crude as this analysis is, the idea behind it is qualitatively valid and the
conclusion is more quantitatively correct than one might expect.
The properties of mirror nuclei also tell us much about the nuclear force.
The data for three A = 27 nuclei are shown in Fig. 1.5 (note that 27 Mg is not
mirror to the other two; it is shown for comparison and contrast). At the right,
the figure gives the relative binding energies or masses of the three nuclei and
on the left the low-lying excited states with their J* values. The remarkable
feature is the nearly identical spectra for the two mirror nuclei. The interac-
tions between two nucleons can be divided into three categories: p-p, n-n, and
p-n interactions. The data from mirror nuclei suggest that the nuclear force is
"charge symmetric" (i.e., the p-p and n-n, interactions are equal). The fact
that the absolute binding energy of 27A1 is greater than 27Si does not reflect a
breakdown of this idea, but rather the influence of the Coulomb interaction:
27
Si has more protons than 27A1 and, therefore, has a greater total repulsive
Coulomb interaction that leads to lower total binding.
Fig.1.4. Highly schematic calculation of the binding energy per nucleon under different "satura-
tion" assumptions concerning nuclear forces. The number of connections indicated , n., is the
number of nucleons with which each other nucleon is assumed to interact. All such interactions
are considered to be of equal strength. The lower part shows a plot of the resulting binding
energies per nucleon.
11
12 Introduction

Fig. 1.5. Level schemes and binding energies (inset box at right with binding energies in MeV
relative to 27A1) of the mirror nuclei27A1 and 27Si, as well as the isobar "Mg.

A more general characteristic of the nuclear force is charge independence,

which means that the p-p, n-n, and p-n forces are equal. To examine this
question, consider an isobar triplet such as 1226Mg14,1326A113, and 1426Si12. The
low-lying levels of 26Mg and MSi are similar as expected: 26Mg has excited states
2+(1.81 Mev), 2+(2.94 MeV), 0+(3.59 MeV), and so on, while 26Si has 2+(1.80
MeV), 2+(2.78 MeV), 0+(3.33 MeV), and so on. At first glance, the nucleus26A1
appears quite different, but careful inspection of its level scheme shows a
subset of excited states with similar binding energies as its isobars. Specifically,
at energies relative to the lowest 0+ state there are 2+(1.84 MeV), 2+(2.93 MeV),
0+(3.52 MeV) levels, and so on. This would seem to suggest that charge inde-
pendence has approximate validity. However, there are other states that have
no analogues in MMg and 26Si. We also note in Fig. 1.5 that 27Mg is quite
different than 27A1 and 27Si, even though Table 1.1 shows that all three A = 27
isobars have the same total number of interactions (351). These results focus
on a crucial aspect of the p-n system: it can exist in two different configura-
tions. The concept of the proton and neutron as merely two states of the same
particle, the nucleon, leads to the concept of isospin, which is similar to
intrinsic spin. In analogy with intrinsic spin, each nucleon is assigned an
isospin t -1/2: protons and neutrons are distinguished by the projection of this
isospin on an imaginary isospin z-axis. This projection tz is -1/2 for protons and
+1/2 for neutrons. Then, just as one can couple the intrinsic spins of two
identical nucleons parallel or antiparallel to S = 1 and 5 = 0, the total isospin
projection of a proton-neutron system can either be T = 1 if the two isospins
are aligned and 7\ = 0 if they are antialigned. Since 7" cannot be less than its
projection, a proton-proton or a neutron-neutron system must have T=l with
TZ = -1 for the proton case and +1 for the neutron case. However, a
proton-neutron system with tz components of -1/2 and +1/2 can couple to
Tz = 1 or 0 and hence T = I or 0. It turns out that the p-n interaction is not
identical in the T = 1 and T = 0 modes. By charge independence, the
interaction in the T = 1 p-n system must be identical to that in the p-p and n-n
 Introduction 13

Table 1.1. Nucleon-nucleon interactions in A = 27 Nuclei

27
A* 27
uMgis 13 Al]4 14 ^13
p-p 66 78 91
n-n 105 91 78
p-n 180 182 182
Total 351 351 351
Based on deShalit, 1974.

systems. However, the interaction in the T = 0 p-n system need not be the
same. (As we will see, it is considerably stronger.) Thus there is no a priori
reason to expect that 27 Mg, with fewer p-n interactions, should have the same
sequence of energy levels as 27A1 or 27Si and, indeed, it does not. Furthermore,
27
Mg is less bound than 27A1, even though it has fewer protons and might,
therefore, be expected to be more tightly bound. The reason is that it also has
fewer T = 0 p-n interactions (see Table 1.1). This is already one piece of
evidence suggesting the T - 0 interaction is stronger than the T = I . The
concept of the different and stronger, p-n interaction in the T = 0 channel will
be of enormous importance later. For example, it determines why the excita-
tion spectra of odd-odd nuclei differ so substantially from those of even-even
nuclei. Moreover, its effects are intimately connected with those of the Pauli
principle since T= 1 corresponds to a symmetric alignment of the two isospins
in the p-n system, while T = 0 corresponds to an antisymmetric alignment.
Nuclear separation energies provide crucial information about the outer-
most nucleons and therefore about certain subtle aspects of the nuclear force
in the "valence" region. As we will see, the most important nuclear model, the
shell model, treats nuclei in terms of individual nucleons that orbit as inde-
pendent particles in a central potential. Each orbit carries certain quantum
numbers and a specific wave function. This is an excellent approximation of
the actual motion except that there are important "residual interactions"
beyond those encompassed by the central potential that must be considered
when dealing with nuclei containing several particles outside closed or magic
configurations. This will be a major topic of discussion in Chapters 4 and 5. We
showed examples of separation energy data in Figs. 1.2 and 1.3 earlier and
summarized the trends, which are valid for all mass regions: that S(p) de-
creases with increasing Z, that S(n) decreases with increasing TV, and that each
increases with increasing number of particles of the other type. Superimposed
on this general behavior is a fine structure in that S(p) and S(n) display
odd-even oscillation in Z and N such that nuclei with even numbers of either
protons or neutrons have larger separation energies (i.e., are more bound).
Though these separation energy data are widely familiar, it is seldom appreci-
ated how much they tell us about the nuclear force.
The separation energies refer to the ground states of their respective nuclei:
in nuclei with even numbers of protons and neutrons the ground state always
has spin and parity JK = 0+. Invariably, this state is much lower in energy than
any intrinsic excitation. The fact that S(p) and S(n) are larger when Z and N
14 Introduction

are even thus implies that there is a special attractive interaction in pairs of
protons or neutrons coupled to JK = 0+. Later, we shall see that this is a property
of short-range interactions resulting from the Pauli principle. The separation
energy data also shows that the p-n interaction is strong and attractive since
S(p) increases with increasing N and S(n) increases with increasing Z. In
contrast, the decrease of each separation energy with increasing numbers of
nucleons of the same type gives the fundamentally critical result that, aside
from the pairing interaction, the residual interaction between like nucleons is
repulsive. This fact, pointed out in the early 1960s by Talmi, is seldom
recognized or remembered; however, its consequences, are profound. For
example, anticipating some concepts and jargon from upcoming chapters, it is
one reason why singly magic nuclei do not become deformed and why the
accumulation of proton neutron interaction strength is essential for the onset
of collectivity and deformation in nuclei.
The properties of the simplest bound nuclear system, the deuteron, tell us
still more about the nuclear force. The essential features, summarized earlier,
are that there is only one bound state, that it has J* = 1+, and that the deuteron
has a finite quadrupole moment. The fact that there is only one bound state
and, moreover, that it is only weakly bound (the deuteron binding energy is
2.23 MeV) serves to emphasize the essential weakness of the so-called strong
nuclear force. By weak here we mean weak in comparison to the kinetic
energy of relative motion of the two nucleons. For example, this implies that
the relative kinetic energy of two nucleons cannot be changed substantially by
the strong interaction. This will be important in Chapter 3, when we discuss
the reason why essentially independent particle motion is possible in a densely
packed, strongly interacting nuclear medium.
It will also be shown in Chapter 3 that for a rather general central potential,
the lowest energy state corresponds to zero orbital angular momentum (an S
state). Thus, both the proton and the neutron in the deuteron must be in S
orbital angular momentum states and the total angular momentum in the
ground state can arise only from the proton and neutron intrinsic spins, 1/2 h
(henceforth, in this book, we shall generally omit the units ft in referring to
angular momentum and intrinsic spin). There are two possible ways of
coupling these two spins: to a total spin S = 0 or 1. The fact that the deuteron
chooses the latter highlights an essential point: even though the nuclear force
may have no explicit spin dependence, there can be large energy differences
between states of different spins in multiparticle configurations. We shall
discuss this point extensively in Chapter 4 where we shall see that the implicit
spin dependence of nuclear forces is a reflection of the Pauli principle and that
this has critical nuclear structure consequences.
Finally, the nonzero quadrupole moment of the deuteron is our first indica-
tion of the tendency of the proton-neutron interaction to lead to nonspherical
nuclear shapes. Moreover, it is an indication that the nuclear force cannot be
completely described by a spherically symmetric central potential. (In the
particular case of the deuteron the finite quadrupole moment is evidence for
tensor forces that couple a spin dependent component to a central potential,
 Introduction 15

but this is not of particular importance in the present context.)

It is worthwhile at this point to reiterate what we have learned about the
nuclear force, and to emphasize that this rather detailed knowledge stems
from some of the simplest empirical facts concerning nuclei. The essential
characteristics of the nuclear force are:
• It is predominantly attractive
• It is short range
• It saturates
• It is charge independent (excluding, of course, the Coulomb part)
Moreover, we have learned that the residual interaction (the internucleon
force not contained within an overall central potential) has the following
properties:
• It exhibits the pairing property that favors the coupling of the angular
momenta of like nucleons to 0+.
• Aside from the pairing interaction, the like-nucleon residual interaction
is, on average, repulsive.
• The T = 0 component of the p-n interaction, on the other hand, is
predominantly attractive.
• The supposedly "strong" nuclear force is strong only in comparison with
other forces: in the nuclear context, it is barely strong enough to over-
come the relative kinetic energies of two nucleons in low energy orbits.
• On balance, as evidenced by the deuteron, the p-n interaction favors the
coupling of the proton and neutron intrinsic spins to 5 = 1 rather than
5=0. In Chapter 4, we shall see that this is in striking contrast to the like-
nucleon residual interaction that favors 5 = 0. Both of these are inti-
mately connected with the effects of the Pauli principle.
• The proton-neutron system has a tendency to produce nonspherical
shapes and provides evidence for spin-dependent tensor forces.
Before we end this discussion of the nuclear force, there is one other
interesting point concerning its range. The short range, ~ 10"12 to 10~13 cm, is
not at all accidental, but may actually be derived by a simple consideration of
its source. It is now generally accepted that all forces in nature result from the
exchange of specific kinds of particles between the interacting entities. Be-
tween the time one of these entities emits such a "virtual" particle and the
other absorbs it, there is a nonconservation of energy. Therefore, by the
Heisenberg uncertainty principle, A; AE > h, there is only a finite amount of
time during which the exchange can occur. Clearly, there is a relation between
the mass of the exchanged particle and the possible range of the force: lighter
(low E) virtual particles induce smaller violations of energy conservation and
therefore can exist for longer periods of time, thus permitting longer-range
forces. The outstanding example of this is the Coulomb interaction, which is
mediated by massless virtual photons and is therefore of extremely long range.
In the nuclear case, the mediation is carried by virtual mesons of which the
lightest are the pions with mass -140 MeV. Assuming that they travel at the
16 Introduction

speed of light, we immediately obtain an upper limit on their "lifetime:"

where m is the pion mass. The distance they can travel in this period is

which is remarkably close to the typical range of the nuclear interaction.

1.3 Paul! Principle and Antisymmetrization

The Pauli principle is of fundamental importance to nuclear structure. For
example, we will see in later chapters that it is essential in determining which
nuclei are stable, that it provides a justification for the idea of independent
particle motion in a dense nucleus, that it is the determining factor in the
energy shifts that occur with various residual interactions in the shell model
and, perhaps most importantly, that it is the principle reason why single
nucleon configuration mixing depends on the valence proton-neutron interac-
tion. In fact, these last two points may seem like structural details, but they
explain in one stroke why all even-even nuclei have 0+ ground states, why the
low-lying states of these nuclei increase in energy with spin, why most low-
lying negative parity states have odd spin, and, remarkably, the entire sys-
tematics of where collectivity, phase transitions, and deformation occur in
nuclei.
The Pauli principle, in its simplest form, embodies the notion that no two
identical nucleons can occupy the same place at the same time. More formally,
no two nucleons can have identical quantum numbers. In this second form it
plays an important role in proton-neutron systems where the two nucleons can
be treated as two states of the same nucleon. Many applications of the Pauli
principle, however, are best expressed in terms of a generalized mathematical
formulation of it that the nuclear wave function must be totally antisymmetric—
totally meaning antisymmetric in all coordinates, spatial, spin, and isospin (i.e.,
that the wave function must reverse its sign if all these coordinates are
interchanged). To see the relation of this requirement of antisymmetry to the
Pauli principle, consider a wave function of two identical particles, v^C^X
where the orbits occupied by the particles are labeled a and b and where ru is
the distance between the two particles. Clearly, the Pauli principle requires
that the wave function must vanish when rn = 0 that is, when the particles are
at the same point in space. A wave function such as ya(r^)yb(r^) need not
vanish at r12 = 0, and thus is not an acceptable two-particle state. However,
consider the wave function

Obviously, y^n) = 0 f°r ri = r2 and thus satisfies the Pauli principle. But, for
any rn, it also follows that
 Introduction 17

So, the Pauli principle can be formulated mathematically by the statement that
a two-particle nuclear wave function *P must be antisymmetric with respect to
the interchange of the two partners. For multiparticle states, the antisymmetry
must extend to interchanges of any pair of particles.
Although the present argument was phrased in terms of spatial coordinates,
it can be extended to other spaces leading to the generalized antisymmetriza-
tion condition given earlier.
It is impossible to overemphasize the importance of the Pauli principle in
nuclear physics. It has obvious and direct consequences as well as subtle,
indirect, but no less real, effects. We shall encounter it continually.

1.4 Two-Slate Mixing

In realistic calculations of nuclear spectra, pure configurations are seldom
encountered. Frequently, the actual nuclear states are complex admixtures of
many components; an accurate treatment must involve the diagonalization of
a large Hamiltonian matrix. Although this is simple, it is tedious and one often
loses sight of the basic physics. In many, if not most cases, however, it is
possible to regain a feeling for the underlying physics and at least a semiquan-
titative calculation by a simple two-state mixing calculation. In many cases,
one can simulate the full diagonalization reasonably well using sequences of a
few two-state mixing calculations. Two-state mixing is completely trivial. We
will present the results in a slightly different form than normally encountered
so that we will obtain universal analytic expressions. It is of the utmost
importance to understand and to have an intuitive grasp of the relationships
between the initial energy spacings and the mixing matrix element, on the one
hand, and the final separations and admixed wave functions on the other.
These ideas are exploited throughout this book. This section outlines the basic
ideas and formulas, presents the universal mixing curve, and discusses some
useful limiting cases. In addition, a few sample schematic multistate mixing
calculations will be described in the next section.
Consider the situation illustrated in Fig. 1.6, in which two initial levels with
energies £j and E^ have wave functions ^ and 02. For an arbitrary interaction,
V, the mixing matrix element is (^IVI^), which we denote simply by V. The
final energies and wave functions are obtained by diagonalizing the 2 x 2
matrix

The final wave functions are denoted by Roman numerals, ¥, and *?„ and have
energies £, and Eu. In general, the mixing depends both on the initial
separation and on the matrix element. A large spacing reduces the effect of a
given matrix element. Conversely, even a small matrix element may induce
large mixing if the unperturbed states are close in energy. In order to present
the results so that this two-parameter aspect is circumvented, yielding a single
universal mixing expression valid for any interaction and any initial spacing,
we define the ratio
18 Introduction

INTERACTION-. v
Fig. 1.6. Two-state mixing: definitions and notation.

of the unperturbed energy spacing to the strength of the matrix element. Then
the perturbed energies are

where the + sign is for En and the - sign for Er It follows that the final energy
difference is

or, in units of the unperturbed splitting AEu, the final separation is given by the
simple result

A more useful result is the amount, AEs , by which each energy is shifted by
the interaction. A/^ I is given by

or, again, in units of A£ u , one obtains a result independent of the initial spacing:
 Introduction 19

The mixed wave functions are

where the smaller amplitude ft is given by

The essential point of Eqs. 1.6 and 1.8 is that both the final energy difference
(in units of A£J and ft are functions only of R, the ratio of the unperturbed
energy splitting to the mixing matrix element. Equations 1.3-1.6 and 1.8 are
universal expressions completely independent of the nature of the interaction
or the initial splitting. The same ratio, R = AEJV always gives the same final
wave functions, energies, and energy shifts (in units of AEu).
These results are so important, and will be referred to, either quantitatively
or qualitatively, so frequently that it is useful to dwell on them. Equations 1.6
and 1.8 are plotted in Fig. 1.7. To illustrate the results and get a feeling for the

hFig. 1.7. Universal two-state mixing curves. The one on the left gives the smaller of the two mixing
amplitudes, ft, while the curves on the right give the energy shift of each level in units of the
unperturbed energy separation. Here the lower curve goes with the upper abscissa scale, while the
upper curve goes with the lower scale.
20 Introduction

Table 1.2. Examples of two-state mixing energy shifts and mixing amplitudes (from Eqs. 1.6
and 1.8). R = AE/V
Specific case: A
R A£/A£B ft V(keV) S ^ '

0.2 4.52 0.67 500 452

0.5 1.56 0.61 200 156
1 0.62 0.53 100 62
2 0.207 0.38 50 20.7
3 0.101 0.29 33.3 10.1
5 0.0385 0.19 20 3.85
10 0.0099 0.099 10 0.99
20 0.0025 0.050 5 0.25
•For R = 0, ft = 0.707, and A£s = V.

numbers involved, let us consider a couple of examples. Suppose two initial

states are separated by 100 keV and admixed with a matrix element of 50 keV
(a not uncommon situation, for example, in Coriolis mixing). Then R = 2, and
we find that the mixing amplitude ft = 0.38 and that each state is shifted by an
amount 0.207 times the initial separation or, in this case, by 20.7 keV. Clearly,
the final separation is 141.4 keV. Another common situation is that of rather
weak mixing. Taking two states initially an MeV apart that mix with a 10 keV
matrix element (R = 100), then Fig. 1.7 or Eqs. 1.6 and 1.8 instantly show that
the mixing is negligible and the energy shift is virtually nil. Table 1.2 gives
examples of AE, and j3 for a range of R values.
Of course, in using these expressions in practical situations one normally
knows the final perturbed energies not the initial separations. One often wants
to extract the amount of mixing (/3) or to deduce the interaction strength from
some experimental measure of the mixing (e.g., the ratio of two transition
strengths from the mixed state, one of which is allowed, the other forbidden in
the unmixed limit: the branching ratio is then directly related to /?). In
principle it is then necessary to work the equations backward to solve first for
V or for AEu, and then for /3. In practice, however, the mixing is often small and
&Es is a small fraction of A£u so that an accurate approximation is obtained by
taking R ~ A£flnal/V - (£n- E^/V. For example, for R > 5, the initial and final
separations differ by less than 10 percent.
Having dealt with some examples and these practical comments, we now
consider two extremely important limiting cases where Eqs. 1.6 and 1.8 sim-
plify: the situations of infinitely strong and relatively weak mixing illustrated in
Fig. 1.8. The results in both cases have many useful and even profound
implications and, for the latter case, the limiting situation has very wide
applicability.
1. Consider first, then, the strong mixing limit. Suppose the two initial
states are degenerate (A£u = 0), as in Fig. 1.8 (left). Of course, then, Eq. 1.6
cannot be used, but Eq. 1.3 tells us that

where E0 is the (common) initial energy. Thus each state is shifted by the
 Introduction 21

Fig. 1.8. The two limiting cases of strong and weak mixing.

mixing matrix element. This illustrates the important result that, for any
isolated two-state system, the final separation can never be closer than twice the
mixing matrix element. As trivial as this sounds, it is often forgotten but is
extremely useful. For example, it was one of the early arguments used to
demonstrate that Coriolis matrix elements had to be attenuated: examples of
isolated pairs of 13/2+ states were found that were closer than twice the
calculated Coriolis mixing matrix elements.
In the case of degenerate orbits, it is clear that /? = I/ V2~ = 0.707. Thus a is
also 0.707 and the two states are completely mixed. This is conceptually
obvious since the matrix element is "infinitely" stronger than the initial
separation (i.e., 1/R —> °°). This seemingly trivial result also has profound
consequences. For example, it means that the mixed wave functions for two
initially degenerate states are independent of the strength of the interaction
between them. (This argument will be used in Chapter 6 to show why the wave
functions in the limiting symmetries of the IBA are independent of the
coefficients—parameters—of the Hamiltonian as long as the structure of that
Hamiltonian corresponds to the symmetry involved.)
2. The weak mixing limit corresponds to R » 1 (see Fig. 1.8, right). Equa-
tion 1.8 becomes

Hence,

since AEs is small. Frequently (for example, from measured y-ray branching
ratios) one has empirical information on /Jand therefore Eq. 1.10 (or the exact
Eq. 1.8) can be used to deduce V from the data. Similarly, for R»l, Eq. 1.6
becomes
22 Introduction

Fig. 1.9. Illustration of thenoncrossingof two admixed levels.

An example is useful. Suppose R = 10. Equations 1.10 and 1.12 then give

The exact results are ft = 0.0985 and A£/A£u = 0.0099. In fact, even for R = 4,
Eqs. 1.10 and 1.12 are already quite satisfactory: /? is correct to belter than 10
percent and AEs to 6 percent. Except in the case of rather strong mixing, Eqs.
1.10 and 1.12 thus provide quite accurate (instantaneous) results for iwo-slate
mixing.
There is one other important aspect of two-state mixing. Suppose we
consider two states, 1 and 2, whose energies depend on some nuclear structure
parameter x (as illustrated schematically in Fig. 1.9). For example, x could he
the deformation and the states might be two Nilsson orbits. For some x value,
xctil, the orbits cross. Now suppose that the two levels mix. They can now never
cross since they repel, and can never be closer than twice the mixing matrix
element after mixing. Thus the actual behavior of the mixed states, labeled 1
and II, is as sketched by the solid lines in Figure 1.9. The energies have an
inflection point. However, for x > jccril, the wave function of state I will have a
larger amplitude for unperturbed state 2 than for its own "parent" and vice
versa. Such behavior is very common in structure calculations and is nearly
always an indication of strong mixing. The point of closest approach of the two
curves corresponds to the point where the mixed wave functions contain equal
admixtures of each of the unperturbed states. In fact, from the separation at
this point the mixing matrix element can be derived by inspection, as one-half
the separation. This is another illustration of the usefulness of the limiting case
ofEq. 1.9.

1.5 Multistage Mixing

In general, a multistate mixing situation must be handled by explicit diagonali-

zation. As noted earlier, this can often be simulated by a sequence of two-state
mixing calculations. In addition, a couple of idealized situations are particu-
larly simple, often useful, and provide physical insight into the often misunder-
stood results of complex calculations (e.g., RPA calculations).
First, let us consider the case of A'degencrate initial states and allow each of
 Introduction 23

Fig. 1.10. Illustration of two multistate mixing situ ations: (Top) N degenerate levels, all of which
mix by equal matrix elements V; (Bottom) The same, except the initial levels are equally spaced.

these levels to mix with each of the others with equal matrix elements (i.e.,
between all pairs). This idea is illustrated in Fig. 1.10. It is then easy to show
by explicit diagonalization that one state is lowered by (N- 1)V and each of the
other states is raised by one unit in V. The wave function for the lowest state
is totally mixed:

Although this is a clear case of (optimum) multistate mixing, the result for the
lowest eigenvalue is exactly what would result from applying a sequence of
two-state mixing calculations: mixing with each of the other W- 1 degenerate
states lowers this state by V, giving a total lowering of (TV - 1) V.
This feature of one state emerging with special character, low energy, and a
highly coherent wave function, is ultimately the microscopic basis for and
physical idea behind the development of collectivity. Collective states result
from many interactions of simpler (e.g., single particle or two quasi-particle)
24 Introduction

entities, and appear at low energies. As we shall see the RPA approach to the
microscopic generation of collective vibrations is a prime illustration of this
effect. So also is the effect of pairing among 0+ states that leads to the well-
known energy gap in even-even nuclei.
A second case is analogous, except that we lift the initial degeneracy and
consider a set of N equally spaced levels. This situation is depicted in Fig. 1.10
for the case of N = 6. As before, one state is considerably lowered. Of course,
the wave functions are now more complex, and are not of particular interest
here. What is interesting is that the ratio of the lowering of the lowest level in
the nondegenerate (ND) case to the lowering in the degenerate (D) case just
considered

is nearly independent of N. For R = 1, L(R) for N = 2,4,8, and 12, respectively,

is found, by diagonalization, to be 0.62,0.60,0.59,0.59: that is the lowest state
is lowered by about 60 percent of what it would be if the initial states had been
degenerate. The near-independence of N means that one can estimate the
lowering, without calculation, simply by taking the two-state mixing result for
the appropriate R value. As a test, suppose the (equal) spacings are all twice
the matrix element V. Then, from Table 1.2, AEs(two-state) is 0.414 V. For the
degenerate case, it is of course (2 - 1)V = V. So L2(R - 2) is 0.414, which should
now be approximately applicable to multistate mixing. The value 1'or N = 8,
obtained by diagonalization, is LH(R = 2) = 0.35.
Finally, a third idealized case again concerns N degenerate levels, except
that each level mixes with only the "adjacent" level (as shown in Fig. 1.11).
This statement, however, is meaningless for degenerate levels, but it is clear
that we can circumvent it by introducing an infinitesimal spacing, and there-
fore an "order" to the unperturbed levels, 1,2, 3...N. This limit, in fact, is not
so far from the realistic situation of Coriolis mixing among a series of bands
with K = Kt, K.+ l.—Kf which frequently occurs in heavy nuclei. Again, one
level is lowered, but now the mixed levels are symmetrically distributed with
respect to the initial energy and the lowest state is not lowered nearly as much.
One can write A£lowcsl =f(N)V where the function/(N) has the rough depend-
ence sketched in Fig. 1.11.

Fig. 1.11. Mullislalc mixing: N degenerate levels in which only "adjacent" levels are mixed w i t h
(equal) matrix elements V.
 Introduction 25

Finally, note that in all the multistate mixing cases considered, all of the
components of the lowest lying wave function have the same sign. Though this
result depends on the phase conventions chosen, if consistent conventions are
used for both wave functions and operators, then matrix elements (observ-
ables) will contain coherent, in-phase sums, and can be extremely large. The
wave function has coherence, and such multistate mixing can lead to collectivity
as reflected in enhanced transition rates, cross sections, and the like. Also,
note that the sum of the initial and final energies is the same, as, of course, it
must be. Since these energies appear on the diagonal of the matrix to be
diagonalized, this is equivalent to the formal statement that the trace is
conserved.
The importance and usefulness of the results in this section cannot be
overemphasized. With them, and an understanding or the basically attractive
nature of the nuclear force, and of the effects of the Pauli principle and of
antisymmetrization, it is possible to understand nearly all of the detailed
results of most nuclear model calculations in an extremely simple, intuitive
way that illustrates the underlying physics that is often lost in complex formal-
isms and computations.

1.6 Two-State Mixing and Transition Rates

One application of the concept of two-state mixing that is worth discussing,
even though it invokes concepts and excitation modes that will be introduced
later, is the effect of certain types of mixing on transition rates. Consider the
simple level scheme in Fig. 1.12 with 2+ levels from different intrinsic excita-
tions (say, belonging to two bands of a deformed nucleus). Suppose that,
according to some model, one 2 f level has an allowed (A) ground state transi-
tion and the other has forbidden (F) transitions to both 0^ and 25+ states. One
occasionally encounters statements of the following kind: "While the 22*—> 0,+

Fig. 1.12. tiffed of mixing on allowed (A) and forbidden (I-') /-ray transitions.
26 Introduction

transition is normally forbidden, its strength results from mixing the two 2+
states: a similar argument accounts for the 22+-> 2* transition." At first, this
certainly sounds plausible: if the two 2+ states mix, some of the strength of the
allowed transition should be "distributed" to the forbidden one. Moreover,
the two 2+ states now share some of the same character and should be
interconnected. Let us calculate the actual E2 matrix elements for the mixed
states to see if the preceding conclusions are warranted. Using the notation of
Fig. 1.6 (Roman subscripts for the perturbed wave functions, arabic for the
unperturbed), we have

since the unpreturbed 1* -> 2* matrix element is forbidden. Thus the

2n+-> O/ transition is now finite and arises solely from the mixing with the 2^
level as claimed. For the 2n+ —> 2 + transition we have

Since the 22+ —> 2* transition is assumed forbidden, the last two terms vanish
and

The 2n+-> 2j+ transition vanishes in the limit of no mixing (j3 = 0). However, it
is by no means clear that mixing will produce a strong transition. The resulting
matrix element is proportional to the difference in the quadrupole moments of
the two states. If the low-lying levels have nearly the same deformation, as is
likely in a well deformed nucleus, this difference will be very small, and thus
the second conclusion is at best risky. We have worked out this example
explicitly because the error just discussed is widespread and partly because the
derivation just given will be useful later in understanding the microscopic
structure of the ^vibration. The point can be summarized as follows: Consider
two states (of the same spin) one of which has an allowed transition to some
other level while the decay for the other is forbidden. The forbidden transition
becomes finite if the two initial states mix, and its matrix element is propor-
tional to the mixing. However, a forbidden transition between the two unper-
turbed levels becomes finite only if the states mix and if the intrinsic structure
of the two unperturbed states differs in the moment corresponding to the
operator for this transition.
With the background provided in this chapter on the properties of the
 Introduction 27

nuclear force, the comments on the Pauli principle and our discussion of two-
and multistate mixing, we can now begin to develop an understanding of the
rich diversity and unity of nuclear phenomena. We start with a cursory survey
of some empirical features and then develop at some length the foundation
models through which we try to understand these data and out of which new
models and extensions arise.
 2
THE N U C L E A R L A N D S C A P E

One of the difficulties often faced by the student trying to understand nuclear
models is that he/she cannot fully appreciate many of the truly simple and
beautiful results that emerge from these models because there is no reservoir
of familiar nuclear data to call upon. Therefore, when one derives the spin
sequence for a 5-function interaction between two identical nucleons in the
same orbit, the results are only of mathematical interest if he or she does not
see that this instantly explains the low lying levels of literally dozens of near-
closed shell even-even nuclei. The entire seniority scheme seems nothing but
a labyrinth of Racah algebra when one does not understand how many well-
known facets of nuclear structure are thereby trivially explained. Similarly,
the simplicity and intuitiveness of many of the results of the Nilsson model may
fall on barren ground unless one realizes the vast number of deformed heavy
nuclei that display exactly these properties.
The main purpose of this chapter is to survey the nuclear landscape to
display a few (definitely not all) typical patterns of nuclear spectra as well as
some of the systematic changes in these patterns over sequences of nuclei, so
that the reader will understand the motivation for each model and will benefit
from an empirical context for their characteristic predictions. We will refer to
the figures in this chapter frequently.
While this approach necessitates some repetition later, it allows us to see
exactly what we are trying to explain with these models beforehand, and what
kinds of data characterize atomic nuclei and are the most useful as tests of
various models.
In principle, at this stage the data should be shown "blindly," without
commentary on its meaning or implications. However, the purpose of this
book is not to develop nuclear physics ab initio and, indeed, most readers will
already be familiar with many of the major concepts and terminology. A
"purist" approach here would be needlessly tedious and artificial. In the pages
that follow, we will use many words and concepts freely that will be introduced
formally later on. Those to whom these concepts are unfamiliar should
concentrate simply on absorbing the data with the idea of using this base as a
touchstone later.
When we speak of nuclear data, we are referring to a vast, varied, and rich
reservoir of information about atomic nuclei—from the deuteron to the acti-
nides—obtained by a most diverse array of techniques. The simplest informa-
28
 The Nuclear Landscape 29

tion is the mass of atomic nuclei. A more useful form for these is nuclear
binding energies, which focus on the interactions between nucleons in the
nucleus when the masses of the individual nucleons are subtracted. Still more
useful (in many cases) are nucleon separation energies, or the energy required
to remove the last, outermost nucleons from the nucleus. (We have discussed
these already in Chapter 1.) The nucleon separation energies give important
data on the surface regions of nuclei. Later, we will show that the individual
nucleons tend to orbit the nuclear center of mass in discrete shells and that, for
many applications, it is possible to neglect all the underlying shells that are
filled. Therefore, the outermost nucleons are frequently most crucial to
understanding the observed properties of nuclear level schemes.
More detailed nuclear data consists of nuclear level schemes: the energies,
and angular momenta, and parity values (J"), of the ground state and low-lying
excited states. The mirror nuclei shown in Fig. 1.5 were our first encounter
with such schemes. Careful measurements of the y-rays emitted when excited
levels de-excite to lower lying ones are fundamental to both understanding
and constructing nuclear level schemes. The crucial information here is, of
course, the y-ray energies, which help define their placements between nuclear
levels and their absolute and relative intensities, which give direct measures of
nuclear transition matrix elements.
A large amount of data has resulted from the study of scattering processes
of one nucleus on another and direct reaction processes in which two interact-
ing nuclei exchange one or more individual nucleons. Scattering experiments
often use low energy projectiles with long wavelengths comparable in size to
the nucleus itself. In such cases, they provide information on the overall
nuclear shape and on the macroscopic, or collective, excitations of the nucleus
as a whole (e.g., rotations and vibrations). Nuclear reactions, examples of
which are single nucleon transfer reactions such as (d, p) or two-nucleon
transfer reactions such as (p, t), can proceed by a direct process in which indi-
vidual nucleons are inserted into or removed from specific orbits. These
reactions provide detailed and microscopic information on the semi-inde-
pendent particle motion characterizing atomic nuclei.
Heavy ion fusion reactions, which can bring in enormous amounts of energy
and angular momentum and access neutron-deficient nuclei, or /? decay ex-
periments on fission product nuclei, which are extremely neutron rich, provide
valuable sources of information, especially on unstable nuclei.
Of course, this listing of techniques barely touches the surface and high-
lights only a few that are most useful for studying low-energy nuclear structure.
A few of these will be discussed in some detail in Chapter 10. For now, we turn
to the picture of the nuclear landscape they have provided us.
The most basic data for nuclei, of course, is a listing of which nuclei exist.
Such a list is usually presented in an JV-Zplot as in Fig. 2.1, where the hatched
area approximately outlines the stable nuclei. This is the so-called valley of
stability (i.e., the valley of energy vs. Z and N: because nature prefers to
minimize energies, the stable nuclei have the lowest energies). The general
features are that N<= Z for light nuclei, and a preference for a neutron excess
30 Introduction

Fig. 2.1. The nuclear chart showing the path of stable nuclei (crossed hatched), as well as the magic
numbers and midshell deformed regions (circles or circular segments).

in heavier nuclei. We shall see that this pattern is easily explained by combin-
ing the concepts of independent particle motion, the Pauli principle, and the
Coulomb force.
We have already seen some other examples of significant nuclear data,
namely binding energies as a function of A, separation energies S(n) and S(p)
and a couple of examples of level schemes for so-called mirror nuclei. These

Fig. 2.2. Low-lying levels of three doubly magic nuclei. Energies on the right of each line are in
MeV.
 The Nuclear Landscape 31

data were shown early on because they provide basic information on the
nuclear force itself, which was already discussed in Chapter 1. Here we are
concerned with a somewhat more detailed and much more extensive survey of
nuclear excitations.
As noted in Chapter 1, there are large gaps in the nucleon separation
energies that define so-called magic numbers, and point to the existence of
nuclear shells, analogous to those in atomic physics. That is, certain magic
numbers of nucleons of a given type correspond to the filling of a set of orbits
constituting a "shell." Additional nucleons must then fill the next higher shell
and are considerably less bound. In Figs. 1.2 and 1.3 we saw such data for the
magic number 82. The complete set of the most important, empirically
observed, magic numbers is 2,8,20,28,50,82, and 126. There are also gaps or
subshell gaps at 40 and 64, especially for protons, which exist for certain
neutron numbers only (see further discussions in Chapters 3 and 6). Figure 2.2
shows examples of nuclei where both the proton and neutron numbers are
magic. Note the extremely high energy of the first excited state, and the
predominance of negative parity states. Figure 2.3 shows the Ca isotopes, two
of which,40|48Ca, are "doubly" magic, while the others are singly magic. The
abrupt change of 2+ energies in 4 2 46Ca compared to 40'48Ca is dramatic evidence
of the difference in their magic structure.
One of the great successes of the independent particle model (Chapter 3) is
the prediction of level sequences for nuclei near closed shells, in particular odd

40 48
Fig. 2.3. Low-lying levels and B(H2) values for the even-even Ca nuclei. Ca and Ca are
doubly magic.
32 Introduction

Fig. 2.4. Low-lying levels of two nuclei, 41Ca and 209Pb, with one nucleon beyond a doubly magic
core, and two other nuclei, 43Ca and 169 Yb, for comparison. Note that the energy scale for Yb is
expanded by a factor of 10. The notations (d, t) and (d, p) indicate which levels arc primarily
populated in these single nucleon transfer reactions. The S(d, p) values given for 209l)b are the
single neutron transfer spectroscopic factors; values near unity indicate nearly pure single-particle
structure.

mass nuclei where, as we shall see, the total angular momenta J (this will often
be colloquially referred to later as the "spin" of the level, although this
nomenclature is clearly inaccurate) of the ground and low-lying excited states
are given by the; values of the orbits into which the last odd nucleon can be
placed. To illustrate this, we show the level schemes of two nuclei (41Ca, 209Pb)
with one particle beyond a doubly magic nucleus in Fig. 2.4. The specific order
and energies of these levels will be easy to understand and predict after we

Fig.2.5. Low-lying levels of three nuclei with two valence nuclcons.

 The Nuclear Landscape 33

Fig. 2.6. Systematics of the lowest levels of the Sn nuclei.

discuss the independent particle model in the next chapter. At this point, these
sequences appear as unintelligible jumbles.
Of course, most nuclei have more than one valence nucleon. Two examples
are included in the figure. One is 2043Ca23 with three valence neutrons. It is
similar in many ways to 4l Ca but, with three valence nucleons, a proper
treatment requires the study of multinucleon configurations and of the "resid-
ual interactions" occurring among nucleons in the valence shell.
The other is 169Yb, which is far from magic in either protons or neutrons.
This scheme, both in terms of its complexity (only hinted at in the figure) and
its compressed energy scale, sets it completely apart from the other nuclei in
the figure. We will see another example of this type of nucleus in a moment.
In treating multivalence particle nuclei such as these, a number of different
approaches are used. Close to closed shells (e.g., 43Ca), an extension of the
independent particle model that includes residual interactions among the
valence nucleons, has been enormously successful. Further from closed shells
(e.g., 169Yb), nonspherical shapes appear, and a deformed shell model (the
Nilsson model) becomes appropriate. In the case of the shell model, it is often
appropriate to consider a coupling scheme in which each nucleon has a given
total angular momentum /. The coupling of these individuals j values leads to
the final J for the state in question. The energy of such a state clearly depends
on the residual interactions among the nucleons in these orbits. It is certainly
one of the triumphs of the shell model that one can easily derive expressions
for these energies, often without a detailed knowledge of the residual interac-
tion itself, that account reasonably well for a large body of data in both odd and
even nuclei.
Figures 2.5-2.7 show the level spectra of some even mass nuclei in the
general vicinity of closed shells. Typical of such nuclei, all have 0' ground
states, first excited levels with ,/" = 2', and mostly even-parity, low-lying
excitations.
All of these features will emerge later from very general considerations of
34 Introduction

Fig. 2.7. Low-lying levels of the A =130 isobars that show the effect of adding valence nucleons (in
this case neutron holes relative lo N = 82).

the dependence of the like-nucleon interaction on total angular momentum

and of the effects of the Pauli principle. Figures 2.5 and 2.6 deal with singly
magic nuclei. They show two interesting features: relatively high-lying, first
excited states and a compression of positive-parity energy levels as J increases.
Both features contrast sharply with "collective nuclei." Note that these
features persist in the Sn isotopes even when there are many valence neutrons.
It is only when there are both valence protons and neutrons that the excitation
patterns change rapidly. Figure 2.7 shows this for three A = 130 nuclei. Note
the systematic change as the total number of valence nucleons increases (here
we count proton particles plus neutron holes relative to the nearest closed
shells). Figure 2.8 illustrates this even more systematically and dramatically.
Here, there is a sharp drop in E 2 f from magic Sn to nonmagic nuclei and the
lowering is greater for more and more valence neutrons. There is an additional
drop when going from two valence protons (Te, Cd) to four (Xe).
In nuclei far from closed shells where the shell model is either intractable or
unreliable, one normally lakes recourse in other theoretical frameworks. One
of the significant and most fruitful of these approaches can be called geomeirleal
or collective models, which bypass the shell model by taking a more macro-
scopic approach of assigning a specific shape to the nucleus and examining the
rotations and vibrations of such a (generally nonspherical) shape. Of course,
a critical issue is whether or not such structures can in fact be derived micro-
scopically from the shell model, and this will be a topic of some importance in
a later chapter. Be that as it may, it is an undeniable empirical feature of many
heavy nuclei that they exhibit properties that seem at variance with the concept
of a shell model, and show evidence of "collective" behavior.
Individual nuclei exhibit several easily discernible types of collective beliav-
 The Nuclear Landscape 35

Fig. 2.8. £2^ values for nuclei in the Sn region.

ior. Figure 2.9 shows typical vibrational nuclei, especially 118Cd, where the first
excited state is a quadrupole phonon excitation of a basically spherical shape.
At about twice and three times this energy there are groups of states that can

Fig. 2.9. Low-lying levels of some typical, near harmonic vibrational nuclei.
36 Introduction

Fig. 2.10. Low-lying rotational and vibrational levels of some typical deformed rare earth nuclei.

be described as two- and three-phonon excitations of the basic spherical

structure.
Figure 2.10 shows three even-even deformed nuclei in the rare earth
region. The lowest levels of spin J = 0, 2, 4, 6,...form a rotational structure
whose energies closely follow the 7(7 + 1) law for a rotating symmetric lop.
Above these are groups of levels, some of which we will interpret in Chapter 6
in terms of intrinsic excitations called ft, 7, and octupole vibrations, each with
rotational bands superimposed. Finally, Fig. 2.11 shows a typical odd mass
deformed nucleus. Here, the levels, which look hopelessly complicated on the

Fig. 2.11. Level scheme of 161Dy. (Left) All levels. (Right) Levels arranged into rotational bands
with Nilsson assignments.
 The Nuclear Landscape 37

Fig. 2.12. £2! values for all even-even nuclei (Raman, 1987).

left, are arranged in sequences of single panicle intrinsic (Nilsson) states, each
with a rotational hand built on top of it on the right. Another example of a
nucleus that can be classified in similar manner is 169Yb, which we looked at in
Fig. 2.4.
We note that the rotational bands in 161Dy range in character from several
examples (labeled 5/2-[523], 3/2-[532], 5/2-[512]) with regular spacings that in-

Fig. 2.13. /?2j values (in keV) plotted against ATor the/I = 100region.
38 Introduction

Fig. 2.14. EI+ values (inkeV) against N for the A = 130 region.

crease smoothly with J, to some with highly irregular sequences [e.g., 5/2+[642],
l/2-[521] and l/2-[530] (with missing 1/2- level)]. An acceptable model for odd
mass deformed nuclei must be able to account for both types of behavior. We
will discuss these intrinsic excitations and their connection to rotational mo-
tion in detail in Chapters 7 and 8.

Fig. 2.15. values plotted against N for the nuclei with N > 30.
 The Nuclear Landscape 39

A dramatic way to illustrate both the collective behavior of nuclei far from
closed shells and the evolution of structure is to examine a particular property
over extended sequences of nuclei—that is, to examine nuclear systematics.
Three of the most telling data are collected in Figs. 2.12-2.16. Figure 2.12
shows the energies of the first excited 2+ states (2 + levels) in even-even nuclei
throughout the periodic table. Figures 2.13 and 2.14 show a more detailed view
of the same data in two particular regions: nuclei around mass A = 100, and
those near A = 130. As we have seen near closed shells, £2+ is rather high
lying, typically 1 to 2 MeV. In contrast, in collective nuclei, the 2t+ state can be
described as either a vibrational or rotational excitation and occurs at much

Fig. 2.16. B(E2:0+1 -> 2^) values for all even-even nuclei. (Bohr, 1975.)
40 Introduction

lower energy. These figures highlight the enormous differences in

E2\ characteristic of these different structures and the transition regions
between them as well as the complexity of some individual regions. Figure
2.15 shows a plot of the energy ratio of the energy of the 4t+ state (first 4+ state)
to the 2* state in even-even nuclei. As will become evident throughout the
subsequent chapters, this ratio is one of the most important structural signatures
and, moreover, is one of the few whose absolute value is directly meaningful.
At first glance, Fig. 2.15 appears to be a semirandom scattering of points.
More careful inspection, and some hindsight from subsequent chapters, shows
that E4+1 E2\ tends to fall into three ranges, values below 2.0 near magic
nuclei (see Fig. 2.5), between 2.0 and 2.4 slightly further away from magic
numbers (see Fig. 2.9), and values very close to 3.33 in midshell regions
corresponding to rotational motion (see Fig. 2.10). The transitions between
the latter two clusters are abrupt indeed.
Most low-lying nuclear states, including essentially all that we will consider,
are below the energy threshold for particle emission and hence de-excite
primarily by electromagnetic processes. These are usually y-ray transitions,
although electron conversion is an important process. The matrix elements for
y-ray transitions often directly give critical nuclear structure information.
We will have many occasions later on to discuss such radiation from excited
nuclear states. This is therefore an appropriate place to deal with the crucial
issue of the angular momenta, A, carried by such radiation and the character-
istic probability of each. As is well known, the electromagnetic field can be
expanded in multipoles of either electric, EX, or magnetic, MX, type. We note
that the parity carried by such photons is (-1)* for electric and (-1)*+1 for
magnetic multipoles. Thus, for example, E2 and Ml transitions conserve
parity while El transitions change it. The probability of emission of radiation
of a given multipole is governed by the intrinsic probability of that multipole
times a nuclear matrix element. The latter depends on the detailed structure of
the initial and final states involved, while the former is a general characteristic
of the electromagnetic field and of the "source" of the radiation (the nucleus).
We will discuss the properties of the transition matrix elements in a number of
different models. Here, we wish to demonstrate in a simple way (ignoring the
intrinsic spin of the photon) that such radiation is nearly always characterized
by low multipoles, or at least by the lowest multipoles allowed by angular
momentum conservation. The argument is very simple. We start by recalling
that the orbital angular momentum l = rxp. The linear momentum carried by
a photon isp = E Ic. Therefore the maximum angular momentum (in units of
K) is given by / = E, Rlhc where R is the nuclear radius. Since fie ~ 200 MeV fm,
dipole (/ = 1) y-rays emitted at typical distances of 10 fm from the nuclear
center must have E - 20 MeV. Since most transitions involved in low-energy
nuclear structure are less than 2 MeV, it is obvious, first, that the electromag-
netic de-excitation process is relatively slow on a nuclear scale (it must proceed
by virtue of the tails of nuclear wave functions extending out to large dis-
tances) and, secondly, that high multipoles (A> 2) are extremely unlikely. Both
these features are very important and are empirically well known. Indeed, it is
 The Nuclear Landscape 41

often assumed that all transitions are El, Ml, or E2, when assigning lvalues,
if the multipolarities have not been measured.
Since the ground state of even-even nuclei is 0+, the first excited state
(normally /* = 2+) can only decay by electric quadrupole or E2 radiation. Since
we shall see that "collective" effects in low-lying states are quadrupole in
geometric character (most deformed nuclei are prolate shaped), it should not
be surprising that E2 or electric quadrupole radiation is of paramount interest.
The usual quantitative measure of E2 transition strengths is called a B(E2)
value, defined as

in terms of the reduced E2 matrix element between initial and final states.
Figure 2.16 shows the systematics of B(E2: 2* -> Oj + ) values across the
nuclear chart in units of so-called single particle values that roughly corre-
spond to the strength expected if a single nucleon changes orbit. The most
obvious feature of the data is the relatively small values near closed shells and
the enormous ones that occur in midshell regions, such as those near mass 160
and 240. These peaks offer the most dramatic evidence known for nuclear

Fig. 2.17. Systematics of /j and 7vibrational energies in the rare earth region.
42 Introduction

collectivity. Within such peaks, there is a characteristic structure: at first, a

rapid increase with increasing valence nucleon number and then a saturation
near midshell. We shall discuss these points, and their implications, in Chapter
6.
It is interesting to make a correlation between E 2 +, £ 4 + / E2+ , and
B(E2:21+-»01+) values. Using our classification of the E^l Ez+ ratios, we see
that low values of E^\IE%\ near closed shell correlate with low values of
B(E2: 2* -> Oa+) and high values of Ez\. As £ 2 j begins to drop as one
proceeds through a major shell, £ 4 + /£ 2 j- rises slightly to just above 2 and the
B(E2) values also begin to increase. Finally, far from magic numbers where
Ei\ drops dramatically and becomes asymptotically constant, Ei,\IEi\ ap-
proaches 3.33 and the B(E2) values increase rapidly toward their peak values.

Fig. 2.18. Systematics of some typical B(E2) values and branching ratios relating to rotational and
vibrational excitations in the rare earth nuclei (Casten, 1988a).
 The Nuclear Landscape 43

As we will study in detail later, the structural transition involved here is one
from spherical nuclei near closed shells (E^lEZ+ < 2) toward spherical, but
vibrational, nuclei and culminates in a phase transition to strongly deformed
(nonspherical axially symmetric) nuclei whose low-lying states reflect rota-
tional behavior. If we recall the discussion of multistate mixing in Chapter 1
(for example, see Eq. 1.13) we see that both the drop in Ei\ and the rise in
B(E2) values can be understood in terms of the correlations or collectivity that
develop from multistate configuration mixing as one goes from closed shell
regions toward midshell. Many of these ideas will be quantified and specific
later.
Finally, a crucial test of our detailed microscopic understanding of the
macroscopic collective shape vibrations is whether we can understand their
systematics. Figure 2.17 shows the energies of/? and /vibrations in the rare
earth region of deformed nuclei. Figure 2.18 gives some crucial B(E2) values
relating to the lowest states in even-even nuclei and to their low-lying /3 and y
vibrations. The top panel shows B(E2: 1 + -> O/) values that describe the
matrix elements connecting rotational states. (This is a small subset of those in
Fig. 2.16.) The middle and lower panels give the ratios of y-vibrational to
ground state B(E2) values to B(E2: 2+-* 0^), and the ratio of "ft -^ g"
over "y-> g" transition strengths. The intraband values are a couple of orders
of magnitude larger than interband B(E2) values and "y-» g" matrix elements
dominate "/? -> g" values. A successful collective model must account for all
these results.
In closing this chapter it is appropriate to summarize the nuclear landscape
in a compact form that will later allow us to make instant, a priori, estimates of
the likely structure of any given nucleus. We do this by recalling Fig. 2.1, which
shows the nuclear chart in an N-Z plane. The magic numbers are indicated by
vertical and horizontal lines and the known and expected midshell regions of
deformed nuclei are encircled. Much of the rest of this book is devoted to
understanding nuclei of each specific type occurring in the chart as well as the
evolution of structure from one type to another.
This page intentionally left blank
 PART II

SHELL M O D E L A N D R E S I D U A L
INTERACTIONS
This page intentionally left blank
 3
THE I N D E P E N D E N T PARTICLE M O D E L

In Chapter I we discussed some of the basic characteristics of the nuclear force,

establishing that, aside from a very short-range repulsive core, it is principally
attractive in nature, rather short range, saturates, and is charge-independent
(excluding, of course, the Coulomb interaction). We also noted that, while the
nuclear force is much stronger than the electromagnetic interaction (indeed, if
this were not the case, nuclei would not be bound), it is nevertheless a rather
weak interaction when compared to the typical kinetic energies of nucleons
inside the nucleus.
In this chapter we discuss the independent particle model, which provides
an indispensable theoretical framework for all that follows. It is the basis for
the multiparticle shell model, which in turn remains the standard of compari-
son for other models and provides the justification, rationale, and microscopic
basis for macroscopic, collective models.
To be clear from the outset, we define some terminology. By independent
particle model we refer to the description of a nucleus in terms of noninter-
acting particles in the orbits of a spherically symmetric potential U(r), which is
itself produced by all the nucleons. Because of this, we immediately anticipate
that the resulting orbit energies are mass dependent. The independent particle
model is applicable in principle only to nuclei with a single nucleon outside a
closed shell and, even then, incorporates certain results from the shell model.
By the latter we refer to a model applicable to nuclei with more than one
valence nucleon that includes residual interactions between these nucleons
and allows for the breaking of closed shells.
Anyone who has studied nuclear physics at all is aware that the basic tenet
of the independent particle model is that the nucleons move essentially freely
in a central potential that is usually taken as a modified harmonic oscillator or
modified square well potential. A little thought, however, raises two appar-
ently serious difficulties before one even attempts this type of aproach to the
nuclear problem. The first centers on the question of how one can validly
speak of independent particle motion in the presence of the strong nuclear
interaction and a densely packed nucleus. The answer involves the Pauli
principle and the essential weakness rather than strength of the nuclear force
referred to earlier. The second question relates to the apparent conflict
between a short-range nuclear force and the nature of a harmonic oscillator
potential that actually becomes stronger as the distance from the origin (the
center of mass) increases. We will return to both issues later.
47
48 Shell Model and Residual Interactions

We saw in Chapters 1 and 2 that the nuclear force is attractive and short
range, and that the systematics of certain gross nuclear properties, such as
nucleon separation energies, are generally smooth, except at certain specific
nucleon numbers, called magic numbers, where they exhibit discreet jumps.
The concept of magic numbers and the shell structure that they imply is of
paramount importance in nuclear physics. Thus we summarize a bit of the
voluminous evidence of their existence. Beside the sharp drop in S(n) and S(p)
just after magic numbers (see Figs. 1.2 and 1.3), the lowest excited states in
nuclei with magic numbers of either protons or neutrons are, on average,
extremely high lying. In particular, in nuclei with even numbers of protons and
neutrons the energy of the first excited state is nearly always a 2+ state, and is
much higher in magic nuclei. This was illustrated by the Ca isotopes in Fig. 2.3:
40
Ca and 48Ca correspond to the magic numbers at 20 and 28. Across the even
^Sn nuclei that have a magic number of protons, the first excited state (2+) has
an energy £2f ~ 1200 keV (see Fig. 2.6) as opposed to E 2 f = 500 keV for the
isotones of Cd or Te (see Fig. 2.8). Even more striking, when Sn becomes
doubly magic at 132Sn, £ 2 + suddenly jumps to several MeV. Further support for
the idea of magicity stems from the fact that elements with magic proton
numbers have higher relative abundance, a larger number of stable isotopes,
and relatively low neutron absorption cross sections. The nucleosynthesis of
such elements predominantly occurs in stellar supernova explosions where an
intense neutron flux leads to rapid, successive neutron capture reactions. This
is the so-called r-process ("r" for rapid). As we shall see in Chapter 10, the
cross sections for such reactions depend mostly on the level density at excita-
tion energies near the neutron separation energy. Such level densities are
particularly low in magic nuclei. Therefore, for magic nuclei, the low neutron
cross sections imply that, once formed, it is unlikely that a sufficient number of
neutron captures take place in the short-lived astrophysical environment to
deplete their numbers. In essence, they tend to block the r-process path.
Thus, we see several lines of evidence pointing to the importance of magic
numbers. Moreover, we notice a relationship in these lines of evidence; many
stem ultimately from the difficulty of exciting a magic or closed shell structure,
and the consequent low-level density at low excitation energies.
Combining all the evidence, we can summarize the relevant magic numbers
for nuclei as

As we shall see in the discussion of nuclear phase transitions, 40 and 64 are in

some cases weakly magic over limited ranges of N and Z.
It is well known in atomic physics that electron binding energies undergo
sharp changes just after a closed electron shell. Analogously, it is reasonable to
suppose that in the nuclear case, these magic numbers correspond to closed
shells of nucleons. Of course, this viewpoint already presupposes a shell model
and we will have to see later whether this provides an apt description of
nuclear properties for nonclosed shell nuclei. Nevertheless, if one wants to
pursue a shell model approach, it is clear that one of the basic features it must
 The Independent Particle Model 49

reproduce is the particular stability of nuclei with these magic numbers. One
would therefore like to construct a nuclear potential that automatically and
naturally produces gaps in single particle level energies at the magic numbers.
It is worth noting an often misunderstood point here: One often hears that
closed shell nuclei are the most stable nuclei. This is not true, however, as a
glance at the chart of separation energies in Figs. 1.2 and 1.3 clearly indicates.
As nucleons of a given type (e.g., neutrons) are added, neutron separation
energies systematically decrease. Just after a closed shell, the separation
energy undergoes a much larger drop. Thus, closed shell nuclei are only more
stable relative to succeeding nuclei.
In considering an appropriate potential for the nuclear case, a tremendous
simplification results if the potential is central, that is, if it depends only on the
radial distance from the origin to a given point. This is equivalent to requiring
that the potential is spherically symmetric. Then, the angular dependence of a
particle wave function is independent of the detailed radial behavior of the
central potential. Moreover, the orbital angular momentum operator, /, com-
mutes with the energy (tf) and is a constant of the motion. All detailed effects
of the particular choice of central potential will therefore show up only in the
radial behavior of the wave functions.
Before considering the specific choice of central potential for the nuclear
case, it is useful to summarize a few general properties of such potentials. We
denote an arbitrary central potential by [/(/•) and only require that U(r) is
attractive and U(r) -> 0 as r —> 0. The Schrodinger equation for such a
potential is

This equation is separable into radial and angular coordinates and therefore
the solutions yT rtm, can be written
n

Here, n is the radial quantum number, / the orbital angular momentum and m
the eigenvalue of its z-component, lf. It is conventional in nuclear physics to
give names to different / values following the convention:

For a given /, m takes the values /, 1-1,1-2...0,-1,-2...-(/-1),-/. Since [/(/•)

is spherically symmetric the (21 +1) energies are independent of m and we will
usually delete this index. The radial Schrodinger equation is

Its solutions have some interesting properties. First, outside the potential, the
wave function decreases exponentially and therefore vanishes as r —> °°. The
quantum number n specifies the number of nodes (zeros) of the wave function
with the usual, but not universal, convention that one counts the node at
infinity but not that at r = 0, that is, n = l, 2,....
50 Shell Model and Residual Interactions

It is easy to prove an important property for such a general central poten-

tial—given two wave functions with different n values and the same / value, the
one with the higher n (more nodes) will have higher energy. Physically, it is
easy to see why this is so since, except for a final exponential falloff, the wave
function must be contained within the range of the potential. Since the
solution with higher n has more nodes, the wave function must "turnover"
more rapidly within the range of the potential. The kinetic energy must
therefore be larger (smaller wave length ).
Similarly, for two states with the same n value, but two different / values,
that wave function with the higher / also has the higher energy. This is also easy
to see from Eq. 3.4, since the centrifugal potential is higher for the particle with
larger /. Therefore this particle has higher transverse motion and is, on
average, further from the nucleus and therefore less bound. These two results,
for the behavior of E^ with n and /, provide the basic reason why shell structure
exists. From these very general and intuitive results one can already deduce an
important conclusion: For any well-behaved central potential, the lowest
single particle state is always an 5 state (/ = 0) and has n = 1. In particular, this
simple result explains why the ground state of the deuteron is primarily an
orbital angular momentum s state. (The small d state admixture is due to
noncentral potentials, which do not have the properties just discussed).
We now turn to the harmonic oscillator potential. This potential is particu-
larly popular in nuclear physics for two principle reasons: It provides a re-
markably good approximate solution to many nuclear problems and it is
particularly easy to handle mathematically, thus yielding many results analyti-
cally. It is given simply by:

The eigenvalues Enl are:

The wave functions, \jfnlm are given by Eq. 3.3.

The detailed specifications of the radial wave functions Rnlare of little prac-
htical importance in the present discussion except to note that they are propor-
tional to Laguerre polynomials in r2 Figure 3.1 shows the form of a harmonic
oscillator potential as well as a square well and a typical modified harmonic
oscillator potential.
The energy levels of the harmonic oscillator potential are shown in Fig. 3.2.
They display two important properties that are evident from the expression for
EM. First, the energy levels fall into degenerate multiplets defined by the
(integer) values of In +1. Secondly, a given multiplet generally contains more
than one value of the principle quantum number n and of the orbital angular
momentum /. A change of 2 units in / is equivalent to a single unit change in n.
Thus, as evident in the figure, the levels 3s, 2d, and Ig are all degenerate.
Physically, this is entirely reasonable in view of the arguments above concern-
ing the sequencing of energy levels in an arbitrary central potential as a
function of n and /. Specifically, since the energy due to centrifugal effects must
 The Independent Particle Model 51

Fig. 3.1. (Top) IlJustration of a single valence nucleon orbiting a doubly magic nucleus. (Bottom)
Schematic illustration of three shell model potentials, a simple harmonic oscillator, a square well,
and an intermediate shape or modified harmonic oscillator. The latter simulates, to some degree,
the effect of an I2 term.

increase with / as well as with the number of nodes (n) in the wave function, it
is clear that, at least qualitatively, one can compensate an increase in n with a
decrease in /. The factor connecting these effects is exactly 2 for harmonic
oscillator potential: it is also 2 for a square well potential.
52 Shell Model and Residual Interactions

Fig. 3.2. Single-particle energies for a simple harmonic oscillator (S.H.O.), a modified harmonic
oscillator with / 2 term, and a realistic shell model potential with / 2 and spin orbit (/ • s) terms.

It is this grouping of levels that provides the shell structure required of any
central potential useful for real nuclei. If we recall that each energy level has
2(21+1) degenerate m states, then, by the Pauli principle, each nl level can con-
tain 2(21 + 1) particles. Therefore, if we imagine filling such a poten-tial well
with fermions, each group or shell can contain, at most, the specific numbers of
particles indicated in the figure. Hence, such a potential automatically gives a
shell structure rather than, say, a uniform distribution of levels.
Unfortunately, except for the lowest few, these shells do not correspond to
the empirical magic numbers. Therefore, while the harmonic oscillator poten-
 The Independent Particle Model 53

tial is a reasonable first order approximation to the effective nuclear potential,

it must be modified to be useful. It was, in fact, the monumental achievement
of Mayer and, independently, of Haxel, Jensen, and Suess, to concoct a simple
modification to the harmonic oscillator potential that enabled it to reproduce
the empirical magic numbers. This step revolutionized forever the subsequent
history of nuclear physics and has led, either directly or indirectly, to essen-
tially all the progress that has been made since. Their achievement, in effect,
is the creation of a realistic shell model. Following their discovery in 1948, the
extensive and very detailed development of this model has led to an elaborate
formalism that provides not only a direct description of many nuclei but also
the microscopic basis for many macroscopic models of collective properties of
nuclei. Since their work, there have been extensive efforts to derive the
nuclear shell model potential from more fundamental data on the
nucleon-nucleon interaction. We shall not concern ourselves with such efforts
here, but shall consider the potential they proposed, discuss it physically, and
here and in the remaining chapters, draw out many of its implications.
It is possible to use some rather general arguments, based on the short-
range nature of the nuclear force, to suggest some plausible modifications to
the harmonic oscillator potential V. Consider a relatively heavy nucleus with
dimensions significantly larger than the range RN of the nuclear force. Then, as
long as a given nucleon lies inside the nuclear surface by a distance greater
than RN, it should be surrounded rather uniformly by nucleons on all sides. It
is screened from the asymmetric distribution that appears at the boundary.
Therefore, it should experience no net force. In other words, the central part
of the nuclear potential should be approximately constant. Thus, from this
point of view, a square well potential might be an improvement on the
harmonic oscillator. Another possibility is to add an attractive term in / 2 to the
harmonic oscillator potential. It is easy to see why this is equivalent to a
flattening of the effective radial shape of the potential. The effects of an / 2
term increase with the orbital angular momentum of the particle. Therefore
high angular momentum particles feel a stronger attractive interaction that
lowers their energies. However, these are precisely the particles that, because
of the centrifugal force, spend a larger fraction of their time at larger radii.
Therefore the addition of an / 2 term is equivalent to a more attractive poten-
tial at larger radii and comes closer to the desired effect of a more constant
interior potential. In fact, it gives a potential intermediate between that of the
harmonic oscillator and the square well. A Wood-Saxon potential has a flatter
bottom than the harmonic oscillator and also produces effects similar to an / 2
term. In the deformed shell model (Nilsson model), that we will discuss in
Chapters 7 and 8, the spherical limit for the single particle energies is explicitly
expressed in terms of such an / 2 contribution.
The relation of the single particle levels produced by a harmonic oscillator
potential, along with the addition of an / 2 term is illustrated in the middle
panel of Fig 3.2, which shows how the In + I degeneracy of the harmonic
oscillator levels is broken as high angular momentum levels are brought down
in energy.
54 Shell Model and Residual Interactions

It is clear that neither of these alternatives yet produces the magic numbers
observed experimentally. It is easy to do so, however, if one introduces a so-
called spin-orbit force. Thus far, we have not discussed the spin quantum
number explicitly. Nevertheless, it is well known that the nucleon, either
proton or neutron, has an intrinsic spin 1/2, and therefore the total angular
momentum of a nucleon in any orbit is given by the vector coupling of the
orbital angular momentum / with a spin angular momentum s = 1/2. With a
spin-orbit component, the force felt by a given particle differs according to
whether its spin and orbital angular momenta are aligned parallel or antiparal-
lel. If the parallel alignment is favored, and if the form of the spin-orbit
potential is taken as V,.t - -Vls (r)l • s so that it affects higher / values more,
then its effects will be similar to those illustrated on the far right in Fig. 3.2.
Each nl level, such as Ig, will now be split into two, Ig9;2 and lg7/2, orbits with
the former lowered and the latter raised in energy. This instantly reproduces
all the known magic numbers.
The absolute strength of the spin orbit force must be substantial (see Fig.
3.2) to produce the correct magic numbers: indeed, the splittings it produces
must be comparable to those between adjacent multiplets of the harmonic
oscillator potential. Since the constant hatoi the harmonic oscillator potential
is found to be h(0 - 41M"3 (e.g., fico^ 8 MeV for medium and heavy nuclei), it
follows that the Vls(r} must attain nearly such magnitudes.
Since the spin-orbit force is an inherently quantum relativistic effect, it is
not as easy to give a physical picture for it as for the relation between an / 2
force and the effective change in the behavior of the central potential just
discussed. It has been shown, however, to arise naturally, and with the correct
sign, from relativistic effects of the nucleonic notion. It is possible, though, to
give plausible arguments for the radial shape of the spin-orbit potential. These
rely again on the notion that, in the interior of the nucleus, a nucleon should
experience no net force. If the spin-orbit force were large in the nuclear
interior there would be a preference for nucleons with spins aligned parallel to
their orbital angular momentum rather than vice versa and therefore such a
nucleon would not be surrounded by an equal number of nucleons with all spin
orientations. This suggests, although it certainly does not prove, that the spin
orbit force is primarily a surface phenomenon. It is therefore customary to
write:

where V(r) is whatever potential is chosen for the central potential itself and
Vls is a strength constant.
It is worth pausing at this point to emphasize the importance of the spin-
orbit interaction. It is not merely a device that ensures the appropriate magic
numbers. Rather, a significant fraction of nuclear structure research in the last
two decades has relied on and exploited the particular consequences of the
spin-orbit force. To see this, it is necessary to refer more explicitly to the
concept of parity. The parity of a wave function \f/nlm = R^r^Y^O, <p) is given
by the behavior, (-1)', of the spherical harmonic Ylm under reflection, since the
 The Independent Particle Model 55

radial wave function does not depend on the sign of r. Thus nritm = (-1)'. For a
multiparticle system of Af independent particles,
and the total parity is

In the harmonic oscillator potential, the particular 2n + /degeneracy led to

shells containing sets of / values differing by even numbers. Therefore all the
levels of a given harmonic oscillator shell have the same parity. The addition
of a modest / 2 term in the potential does not alter this qualitative fact but a
spin-orbit potential can lower the energy of the; = 1+1/2 orbit sufficiently so
that, at least in the higher shells, it is brought down among the levels of the next
lower shell. Indeed, this is precisely the effect required in order to reproduce
the magic numbers. For example, the positive parity lim orbit now appears in
the 82 to 126 shell among the negative parity 21^, lh9/2, 2f5/2, 3p3/2 and 3p1/2
orbits. Thus, a real shell, bounded by the magic numbers, contains a majority
of levels of one parity and one level of the opposite parity. It is conventional to
call these the normal parity orbits and the non-normal or unique parity orbits,
respectively.
The significance of this will only become clear when we consider certain
residual interactions experienced by the nucleons in the shell model potential.
However, the general point can be made easily. In some cases, these residual
interactions are diagonal, providing contributions only to the energies of the
various levels. In others, however, they have important nondiagonal effects
that mix different configurations. Whatever mixing does occur, however,
cannot mix levels of different parity. (We neglect here the very weak parity-
nonconserving part of the weak interaction.) We recall from the discussion of
two- and multistate mixing in Chapter 1 that such mixing effects are strongly
dependent on the energy separation of the mixed states. Since the unique
parity orbit cannot be admixed with its neighbors in a given major shell and
because it is so far separated from the other orbits of the 2n + 1 multiple! from
which it originated, it mixes only very weakly with other levels. Therefore,
configurations stemming from the unique parity orbit are particularly pure
even though they occur amid an enormous complexity of mutually admixed
states characteristic of many heavy nuclei. These levels therefore provide an
ideal laboratory for testing various nuclear models, since one deals with
particularly pure, simple, and well-known wave functions. This fact has been
exploited in countless experimental and theoretical studies in recent years.
Perhaps the most well known involve the interesting physics of high spin states
in deformed nuclei. Indeed, it should not be surprising that the first, and often
the majority of tests, of virtually every new approach to the study of high-spin
states deals with unique parity levels. Only after whatever new effects are
involved are well understood for these levels does one generally dare to look
at the normal parity levels.
Unique parity levels are also of particular importance for low-spin states
and for reasonably light nuclei. In the next chapter we shall study the effects of
56 Shell Model and Residual Interactions

residual interactions on nuclear level schemes. In general, these are quite

complex and can lead to significant configuration mixing. However, we shall
see that many of the complexities are eliminated, and indeed, rather simple,
analytic, physically-reasonable results emerge, if the nucleons are restricted to
occupying a "single;'" orbit. Clearly the testing ground par excellence for such
ideas, will again be unique parity orbits. The fact that these are also the highest
j orbits in a given shell and thus can contain the most particles, further enriches
the phenomena that can be studied.
Now that we have outlined the basic features of the nuclear potential and
seen the single particle energy levels to which it leads, we are nearly ready to
consider the predictions of this model for various nuclei. The basic idea is
simply to fill the levels of the potential sequentially, as one adds nucleons in
going from nucleus to nucleus. With the help of some simple arguments we will
see that we are immediately able to make many predictions concerning the
low-lying levels of a number of nuclei.
One caveat is appropriate here. When looking at a sequence of levels in a
central potential, as in Fig. 3.2, the overall scale is illustrative only and in fact
changes with mass according to the well-known expression fica- 41/.A"3; there
is a gradual scale compression for heavier nuclei. Moreover, residual interac-
tions affect the single particle energies themselves, as we shall discuss later.
Thus, such sequences are a guide only. In practical use, specific single particle
energies for each mass region must be used.
Before applying the model there are several rather profound issues, seldom
thought about explicitly, that must at least be mentioned. The entire concept
of "filling the levels... sequentially" is in fact one of the most important
applications of the Pauli principle. Without this dictum, any number of
nucleons could go in the lowest orbit. That being the case, there would be no
reason (energetically) to favor the addition of a neutron or a proton (neglect-
ing the Coulomb force—see the following). A nucleus with Z = 2, N = 300
would be as stable as any other. It is one of the most significant consequences
of the Pauli principle that stable nuclei have N** Z since a large excess of either
would involve the filling of higher-lying levels at an extra cost in energy.
Indeed, the entire shape of the valley of stability (Fig 2.1) results from the
countervailing effects of the Pauli principle that favors N=Z and the Coulomb
repulsion that favors a large neutron excess. For low Z, the former effect
dominates, while for higher Z the valley of stability curves toward the neutron
rich side (see the next paragraph). It is the same effect of the Pauli principle
that determines nuclear radii: without it (and the Coulomb force), all nuclei
would be comparable in size to an a particle.
This discussion leads naturally into a few comments about the Coulomb
force, which is simply an extra, repulsive, long-range potential (V^^ ~ I//") felt
by the protons. We will not discuss this in detail, but it is obvious that its
principle effect is to raise the single particle proton energy levels. As we have
just seen, in the absence of the Coulomb force, proton and neutron energy
levels would fill with equal likelihood, there being no distinction between the
two, and the valley of stability would correspond to N = Z = AI2. However,
 The Independent Particle Model 57

because of the Coulomb interaction, there is an extra cost in adding additional

protons. This effect is clearly more important for heavier nuclei since the
Coulomb force is a long range interaction that scales roughly as Z(Z -1), the
total number of interacting proton pairs. Therefore, as A increases it becomes
energetically preferable to add additional neutrons, and thus stable heavy
nuclei have a neutron excess.
A second issue of importance concerns the very concept of independent
particle motion in a central potential. As noted earlier in this chapter, this
seems to be at odds with the idea of a strong nuclear force. Empirically,
nuclear radii can be described by the relation R - 1.2Am. If we take 1 fm as the
radius of a nucleon, and imagine the nucleus to be uniformly filled, then the
ratio of the nuclear volume to that of a single nucleon is just (12)3A, or hardly
more than A times the volume of each nucleon! Therefore, it would seem
unlikely that an individual nucleon could execute countless undisturbed orbits
in such a densely packed medium. (Note that this is in stark contrast to the
emptiness through which atomic electrons follow their orbits. Bohr did not
need to address this issue, nor could he, since its solution is yet another
consequence of the Pauli principle.)
The resolution of the paradox rests on the conjunction of the Pauli principle
with the essentially weak nature of the attractive nuclear force where, by
"weak," we mean relative to typical kinetic energies of nucleons within the
nuclear volume. This weakness is illustrated by the fact that in the deuteron,
the simplest nuclear system, the attractive nuclear interaction is only suffi-
ciently strong to produce one bound state.
To understand the possibility of independent particle motion, consider a
central potential with various energy levels, defined by the quantum numbers
nl (we ignore spin for this argument), in which the nucleons, either protons or
neutrons, sequentially fill each level. Since the central potential is spherically
symmetric, the energy levels are independent of the magnetic quantum number
m and therefore, each such orbit is 21+1 degenerate (for any /there are 21+1
magnetic substates). By the Pauli principle, which states that no two fermions
can occupy the same configuration, such an orbit can contain at most 21+1
particles. Therefore, as more and more nucleons are added to the nucleus,
orbits of successively higher n and / values, and therefore higher energies, are
filled. Now, imagine a pending collision between two nucleons in relatively
inner orbits. By the Pauli principle, unless the impact of the collision is
sufficiently strong to raise one of these nucleons to an unoccupied orbit (i.e.,
above the Fermi surface), the interaction can have no effect on the motion of
these two nucleons. Therefore only the outermost nucleons are likely to be
affected by such collisions. Thus, just as it is the Pauli principle that prevents
the nucleus from collapsing on account of the attractive nuclear potential, it is
also the Pauli principle that leads to independent particle motion for most
nucleons in the nucleus. This does not mean that there cannot be correlations
in two-body nuclear wave functions, but those correlations must be consistent
with the Pauli principle and with the strength of the attractive nucleon-nucleon
interaction. Later, we shall see later that many of those correlations that do
58 Shell Model and Residual Interactions

occur are an implicit effect of the antisymmetrization of nuclear wave func-

tions associated with the Pauli principle.
Another question worth mentioning is inherent in the use of a harmonic
oscillator potential in the shell model: How does one reconcile the idea that a
nucleon at large distances from a real nucleus cannot experience a nuclear
force with the fact that a harmonic oscillator potential increases in strength
with radius? This point is rather subtle and relates to the problem of separat-
ing the internal motion of the nucleons from that of the nuclear center of mass.
This problem has been elegantly discussed in de Shalit and Feshbach, and we
will not dwell on it here except to comment that in a typical nuclear potential
V, such as the harmonic oscillator, where V -» °°, the falloff of nuclear wave
functions with increasing distance from the origin behaves rather differently
(as e~ar2) than is the case for quantum mechanical tunneling through a finite
potential (e -ar). However, this difficulty is primarily a "long distance" prob-
lem: the two wave functions are similar within the nuclear volume. Significant
errors may accumulate in studying the tails of nuclear wave functions. Such
effects can be important, for example, in studying single nucleon transfer
reactions that occur primarily in "grazing" collisions at the nucleus surface.
As a final preliminary, we must consider the structure of closed shells and
their effects on the valence particles. Consider a system of A nucleons (either
protons or neutrons, which, for the present purposes, we treat independently:
the interactions between them will be considered later). Each orbit of spin
j = I ± 1/2 has 2; + 1 degenerate magnetic substates. By the Pauli principle,
once 2j + I nucleons are in a given orbit, it is filled (closed) and the next one
begins to fill. To predict the spins and parities (J" values) of low-lying levels in
nuclei, we need to consider the effects of the (filled) closed shells. First, we
show that the total angular momentum of a closed single-/ shell is identically
zero. This is trivial. It is immediately clear that the total magnetic quantum
number M of a closed shell is M - Imi=j + (j-1)+...+ (- (/-I)) + (-/) = 0. Since
this is the only possible M value, it follows that the total angular momentum
7 = 0 since J can never be larger than the largest M. Alternately, for each state
/, m there is a state/, -m, and therefore the vector sum J = j + j vanishes. Since
this is true for each pair of states (with ±m values) and, of course, for m = 0, it
is true for the full set of 2/ + 1 states. Furthermore, since it is true for a given
/ value, it must be true for an entire major shell consisting of several / values.
Note the importance of this result: without it, it would be impossible to
apply the independent particle model to any nuclei except hydrogen or a
neutron where, of course, the idea of the central potential itself would not be
valid. With it, however, one can at least predict the ground state spins and low-
lying excited states of any nucleus consisting of one particle, either proton or
neutron, outside a closed shell. Since the core—that is, all closed shells (major
or single -/)—contributes only J" = 0+, the angular momentum of the ground
state of a nucleus with one particle in a shell nlj, is therefore just / = / with parity
*=(-!)'.
Any program of predicting ground state J values, or the J values and
energies of single-particle excited states, presupposes that we know the order
 The Independent Particle Model 59

and spacings of the single-particle levels. The reader may perhaps think that
we have already dealt with and solved this question by the sequence on the
right of Fig. 3.2. However, this ignores the fact that the single-particle energies
themselves depend on the number of nucleons in the nucleus since the single-
particle potential arises from these same nucleons. Considering nuclei with
one valence particle (beyond closed shells), this means considering the effect
of the closed shells.
The reader has undoubtedly encountered the idea that closed shells can be
ignored. There is considerable truth in this, as we shall see, but it is not the
whole story. Especially in recent years, with the advent of much new data far
off stability, this issue takes on real importance. In the next few paragraphs we
discuss the effects of closed shells on single-particle energies. An extension to
multiparticle configurations will be discussed early in the next chapter.
The basic result is absolutely trivial. Since a closed shell has 7 = 0, its wave
function is spherically symmetric. Therefore, imagine a single valence nucleon
outside this shell in an orbit; and magnetic substate m. Since the closed shell
has no preferred direction in space, its interaction with this nucleon must be
independent of m. This does not mean that the interaction can be ignored. It
can, and does, exist but it is only equivalent to a change in the spherically
symmetric central potential. A particle in a particular valence ;' shell certainly
interacts with the closed; shells below it and its single-particle energy is altered
by that interaction. The preceding argument simply means that this interac-
tion is independent of direction. [Incidentally, to anticipate our later discus-
sion of the multipole expansion of an interaction in Pk(cos&), this shift is due
exclusively to the monopole (k - 0) part of the interaction between the closed
and open shell nucleons, since that is the only multipole that is 9 independent
(P0(cos0) = constant)].
This interaction of an open shell with underlying filled shells has an impor-
tant consequence. A major shell generally consists of several constituent;'
shells. Each of the ;'s of a closed major shell can have a different (spherically
symmetric) effect on each of the valence ;' orbits. Thus, the relative single-
particle energies in a given major shell depend critically on the specific lower-
lying, filled closed shells.
We can illustrate this effect and understand its importance with some
examples. Empirically, it is possible to map out single particle energies with
one-nucleon transfer reactions. These will be discussed extensively for de-
formed nuclei in Chapter 8. Suffice it to say here that in a reaction like (d, p),
illustrated in Fig. 3.3, a neutron is stripped off the incoming deuteron into a
specific orbit around the target nucleus, leaving an outgoing proton. Clearly,
such a reaction can disclose sequences of states with single-particle structure,
each corresponding to different;' orbits. Moreover, once the "kinematical"
aspects of the reaction collision are removed, the cross sections yield a nuclear
matrix element, which is the purity of the single-particle state. This informa-
tion is embodied in the so-called spectroscopic factor: S(d, p) = 1 corresponds
to a pure single-particle neutron wave function coupled to the target nucleus.
While a reaction like (d, p) can only populate a state to the extent that the
60 Shell Model and Residual Interactions

Fig. 3.3. Schematic illuslration of a direct (d, p) reaction.

corresponding orbit was empty in the target, "pickup" reactions like (d, t)
extract a neutron from an already filled orbit: they produce hole excitations
and sample single-particle energies below the current valence shell.
Extensive data collected over several decades, much of it from such reac-
tions, has allowed us to map out the empirical single-particle energies in a
number of nuclei that are one particle or hole removed from various major
closed shells. The recent extension of such data to nuclei far off stability has
greatly expanded the overview of single-particle energies thus provided.
We summarize some of the results in Fig. 3.4. Each panel gives the observed
single-particle or hole energies for two nuclei, along with schematic illustra-
tions of the orbits involved. The point is to compare these energies for
different systems, cores, and types of nucleon.
The top left panel of Fig. 3.4 shows proton and neutron single-particle
energies in the 82 to 126 shell extracted from the particle levels of 209Bi and the
hole states of 207Pb. The energies are nearly identical. (The slight expansion of
the proton energy scale is probably a Coulomb effect.) This is reasonable. In
each of the nuclei, the valence particle or hole "feels" interactions with
essentially the same underlying orbits. The only notable difference is that the
neutron holes in 207Pb interact with 43 other neutrons in the 82 to 126 shell,
while the proton particles in 209Bi interact with the full major shell of 44
neutrons. As individual nucleon-nucleon interactions are on the order of a
few hundred ke V, and ^'-dependent differences considerably less, this should be
a minor effect. On the top right panel in Fig. 3.4, a similar situation is shown for
the 50 to 82 shell with identical results.
However, in the bottom panel, the neutron holes in 131Sn interact with the
same 50 neutrons and 40 protons as in Zr but, in addition, with the Z = 40 to 50
closed proton shell and with other neutrons in their own shell. As a conse-
quence, their energies are very different from 91Zr.
This particular example is actually of great structural consequence. (We
 The Independent Particle Model 61

Fig. 3.4. Changes in single-particle energies in different nuclei that illustrate the effect of closed
shells on the valence orbits. Since we are interested in relative changes only, one energy is
normalized in each box.

will see later that it accounts for the onset of deformation near A = 100.) The
point is that the interaction of a given shell,;',, with another (closed or open),
jv depends on the overlap of the respective wave functions; the only difference
if j2 is closed is that, then, the interaction is angle independent, so it is the radial
overlap that counts. Orbits with similar quantum numbers nlj have higher
overlaps. This is demonstrated in Fig. 3.5, where the dependence of the radial
overlaps of various orbits with a Is orbit on An = nl - n2 and A/ - l^ - 12 is
illustrated. The falloff with Arc and A/ is clear. (Incidentally, the simple
estimate that the interaction goes roughly as l/(An + A/) is not a bad guide).
In this particular case, as the proton Ig^ orbit fills from Z = 40 to 50 it exerts
a strong attractive pull on the lg7/2 neutron orbit, drastically lowering the
energy as seen so clearly in Fig. 3.4. In contrast, an orbit like 3s1/2 has poor
overlap with lg,J/2 and, relatively, its energy increases.
62 Shell Model and Residual Interactions

Fig. 3.5. Indication of the dependence of a residual interaction on the difference in principle (n)
and orbital angular momentum (/) quantum numbers of the two orbits involved. The contours
give constant values of the radial overlap integrals of a ls1/2 orbit with orbits of different n, I
(Heyde,1987).

The sensitivity of relative single-particle energies in a given shell to the

occupation of different underlying shells is further illustrated in the same
panel by the extension to 207Ti. Here, one observes the proton orbits in the 50
to 82 shell, but now an additional neutron shell (82 to 126) has been filled.
Since this extra neutron shell has, on average, higher;' values (1/2 to 13/2
instead of 1/2 to 11/2 for 50 to 82 or 1/2 to 9/2 for 28 to 50), the main effect is to
further lower the highery-proton single-particle energies (7/2,11/2) relative to
the lower ones (1/2, 3/2, 5/2).
This rather lengthy discussion is important because it highlights the three
points that (1) a closed shell has the same effect on all magnetic subsidies of a
particle in an open shell, (i.e., it is equivalent to a change in the spherical
potential), (2) the energies of different] shells may be very differently affected
by underlying shells, and (3) different underlying closed shells (/ or major) may
affect a given open shell very differently. It also serves to re-emphasize that an
illustration of the "shell model levels," as shown in Fig. 3.2, can only give a
semiquantitative guide: the energies are mass dependent.
With this background, we now turn to actual predictions of the independent
particle model. Despite the caveat just made, for convenience, we will use Fig.
3.2 as a single-particle energy reference. (Note: in most of the following
discussion we will denote orbits by the /;' quantum numbers alone, as in f7/2,
since omission of n simplifies the notation and seldom causes confusion.)
This simple model works extremely well for a large number of nuclei. For
example, by inspection of the right-hand panel of Fig. 3.2, we would expect the
ground state of 2041Ca21, which has one neutron beyond the N = 20 closed shell,
to be 112-. Similarly,2141Sc20 should have a 7/2~ ground state, while the ground
states of 4191Nb50 and 4091Zr51 should be 9/2+ and 5/2+, respectively. These
predictions are verified experimentally.
It is also trivial to predict excited states in the independent particle model:
these can be formed simply by elevating a particle from the ground state orbit
 The Independent Particle Model 63

Fig. 3.6. Low-lying single-particle levels of 209Bi.

to a higher level. The excitiation energy is given by the energy differences

between the orbits in a shell model potential such as that in Fig. 3.2.
For example, in 41Ca, there should be a low-lying 3/2 state and slightly
higher 1/2- and 5/2- states. Another means of forming an excited state is to
elevate a particle from a lower, filled level. Suppose an s1/2 nucleon is raised to
the tj/2 orbit. In principle, we now have to deal with a "hole" in the N = 8 to 20
shell and two particles above N = 20. Anticipating a result from the next
chapter, while the two lf7/2 particles can couple their angular momenta to form
several J values, the J* = 0+state lies the lowest by far for any realistic interac-
tions between these particles. Therefore the three-particle state I (s1/2)-1(f7,2)2-/)
(to use an obvious notation where exponents count particles and negative
exponents count holes), has J" = l/2+ as its lowest state. We thus expect a l/2+
level in 41Ca at an excitation energy given roughly by the £; 7 , 2 -£ nfe orbit
64 Shell Model and Residual Interactions

energy difference. Although this estimate neglects some residual interaction

effects, it is good enough for the present purposes.
The 41Ca level scheme thus should exhibit low-lying 7/2-, 3/2-, 3/2+, and l/2 +
levels. The empirical scheme was shown in Fig. 2.4. Clearly there are many
more levels than we have accounted for. This is not exactly a discrepancy, but
rather a caveat, since the extra states arise from more complex configurations
and can be understood by techniques developed in the next two chapters. In
Fig. 2.4, we indicated those levels identified in (d, p) or (d, t) and we can see
that this "single particle" subset of states in 41Ca is exactly that demanded by
the independent particle model. While this confirms some basic predictions of
the model, the wealth of other states hints at the greater complexity to come.
In the same way, the single-particle states in 209 Pb and 209Bi and many other
nuclei have been identified. For example, the results for proton levels in the 50
to 82 shell and for neutron levels below the magic number 82 were shown for
133
Sb and 131Sn, respectively, in Fig. 3.4. Those for the particle excitations of the
Z = 82 to 126 shell are given for 209Bi in Fig. 3.6. In all three cases, the expected
ordering and spacing is approximately observed. The results for 209Pb arc
included in Fig. 2.4, along with their spectroscopic factors, S(d, p). (The S(d, p)
values have a maximum theoretical value of unity: those empirical values
slightly exceeding unity may be due to experimental errors or difficulties in
removing reaction kinematics from the observed cross sections.) Except for
the 15/2" levels, the states have remarkable purity and display an order in basic
agreement with Fig. 3.2.
Notice that all of the nuclei considered so far have only one nucleon outside
a doubly magic nucleus. In condidering excited "hole" states in 41Ca, we have
already encountered three-particle states and used a result from Chapter 4 for
the relative energies of states in a |;V) configuration. By generalizing this with
a further conclusion from Chapter 5, we can greatly extend the usefulness of
the independent particle model. We stated earlier that it is energetically favor-
able for two particles in an orbit to couple their angular momenta/ to./ = (). As
discussed in Chapter 5, this tendency persists for any even number, n, of
particles in an orbit/, that is, in the configuration/". Therefore, the lowest state
of a /"11 configuration can be thought of in the form /"(.^ = ())/,./ = /> where ./,
is the "intermediate" ./ value for n nucleons. Thus, again ./ = / of the last odd
nucleon; this now allows us to predict sequences of ground and excited state,/
values for odd mass nuclei relatively far from closed shells.
For example, from Fig. 3.2, the odd neutron in the Ca isotopes from 4l Ca to
47
Ca should fill the l'7/2 orbit above the N = 20 magic number. These nuclei
should all have J* = 7/2- ground states. The l'7/2 orbit is filled by 48Ca and the
ground state of 49Ca should have an odd neutron in the p3/2 orbit.
If we consider 3739Ca, we are dealing with holes in the N = 20 neutron shell
(i.e., the d3;2 orbit). It is easy to show that a state with one hole in a / shell also
has/=/. Since 2/mustbeodd, the first (2/-1) nucleons will "pair" off to./ = 0,
using our previous argument, leaving one particle in the orbit/, and hence,
giving ./ = /. Another proof is simply the statement that since the (2/ + l)-
particle closed shell has./ = 0, the only angular momentum in the system of a
 The Independent Particle Model 65

Fig. 3.7. Ground state J* values for a number of odd-mass nuclei for comparison with the
predictions of the independent particle model (compare Fig. 3.2).

one-hole state is the hole angular momentum j. (Implicitly, of course, we

already used this result in predicting the J* values of 47Ca, another one-hole
nucleus.) Thus the ground state of 39 Cashould have7"= 3/2'. In 3 7 Ca,lhe first
two holes pair to 7 = 0+ and the third odd particle again gives the./ = 3/2T ground
stale spin. These predictions are verified by the empirical ground stales in Fig.
3.7 (top row).
This figure shows several other empirical sequences illustrating the predic-
tions of the independenl panicle model. The nucleus 4090Zr50 is doubly magic.
Isotopes with N>50 should be filling the neutron dM orbit until N = 56. Below
N = 50 there will be holes in the g9/2 orbit. So we expect ground state spins
7 = 9/2+ for 87> 89Zr and 5/2+ for 91 ' 93 ' 95Zr. The next line in the figure shows the
odd proton yjY isotopes. Note lhal they all have a l/2~ground state. This is the
expected orbit for the 39th proton. The interesting point here is that the 1/2
level remains the ground state, regardless of the number of neutrons over a
very wide range of nuclei. The 41 Nb nuclei in the next line exhibit the same
feature, except the 41st proton is in the g9/2 orbit. (Interestingly, the first
excited state in each of these Nb isotopes is 1/2 , which can be obtained by
66 Shell Model and Residual Interactions

elevating a proton from below the N = 40 closed shell to pair off with the odd
g,^ proton, leaving a l/2~ hole.)
As a final illustration, if two of the three valence neutrons in 43Ca couple to
J = 0, its excited states, identified in (d, p) or (d, t), should resemble those in
41
Ca. Figure 2.4 confirms this.
From our earlier discussion of Fig. 3.4, it is obvious that deviations from Fig.
3.2 must occur. The bottom row of Fig. 3.7 is a sequence of N = 50, odd Z nuclei
from Z = 35 to 45. From Fig. 3.2 we would expect the 35th and 37th protons to
fill the f5/2 orbit with a low-lying pM excitation, the 39th to occupy the p1/2 orbit
and the 41st to 45th the g9/2 orbits. The predictions for Z = 39 to 45 are
confirmed empirically, but there is an inversion of the fs/2 and pM levels. As
another example, the level order ordinarily associated with the 50 to 82 shell is
a low-lying pair of g7/2, d5/2 orbits with a small gap before a cluster of s1/2, d3/2 and
hn/2 orbits. This pattern is seen clearly in 131Sn in Fig. 3.4. However, this figure
also shows the very different order in 91Zr. While we are able to predict the
d5/2 ground state for Zr, Fig. 3.2 would have suggested a 7/2+ first excited state,
instead of the l/2+ state observed empirically. Indeed, "Zr (with filled d^
orbit) also has a l/2+ ground state. The inversion of the ground and first
excited states of 209Bi is a third example (Compare Figs. 3.2 and 3.6). These are
not deficiencies in the model, just a fact of life in its use. As discussed in the
context of Fig. 3.4, one must deal with changes in single-particle ordering,
especially when the same major shell is inspected in nuclei that are far apart in
mass so that the interactions of the intervening nucleons will have altered the
single particle potential. Also, because of the Columb potential, proton and
neutron single-particle sequences are expected, and found to be, slightly
different, especially in heavy nuclei.
Of course, there are also real discrepancies with the predictions of the
independent particle model. Figures 2.4 and 2.11 showed the examples of
169
Yb and 161Dy, where a totally different empirical picture, completely at
variance with Fig. 3.2, is observed. We will see the reasons for this later.
There is one last prediction of the independent particle model that accounts
nicely for a characteristic aspect of doubly magic nuclei. Clearly, any excita-
tion of a nucleon to the next higher shell nearly always involves excitation into
an orbit with the opposite parity. Therefore, the low-lying excited levels of
such nuclei should be negative parity. This is empirically well known, as
exemplified in Fig. 2.2.
Although we will not discuss it here, there are many other predictions, for
example magnetic moments or single-particle electromagnetic transition rates,
that are reasonably well reproduced by the simple independent particle
approach.
Finally, we caution that, while / = 0+ is the lowest-lying configuration for the
j" configuration of identical nucleons, other / values are possible at higher
energy. Empirically, the next energy level nearly always has J = 2*. Therefore
it is risky to extend the predictions for odd mass nuclei above excitation
energies comparable to Eif in the neighboring even-even nucleus, since con-
figurations of the form j ® 2+, J) in which a particle in orbit nlj is coupled to a
"core" excitation (the first 2* state) of the under-lying even-even nuclei can
then compete with single-particle excitations.
 4
THE SHELL MODEL:
TWO-PARTICLE C O N F I G U R A T I O N S

In order to proceed, we must deal more deeply and systematically with the
problem of multinucleon configurations. By this we mean "valence" configu-
rations of two or more particles outside a core, which is usually assumed to
consist of inert closed shells. There are really two issues here: First, which J
values are allowed by the Pauli principle, and, second, what is their relative
energy ordering. The first question can be answered without reference to any
discussion of a central potential or residual interaction, while the second
depends on the details of those interactions, although many properties can be
deduced from the simple fact that the radial and angular parts of a wave
function in a central potential can be separated and that predictions for the
latter are independent of / value, and for the radial part, many results can be
obtained by rather general arguments based on its short- or long-range charac-
ter or by considering a multipole expansion of it. This chapter addresses these
issues. We will start by considering the simplest case of two identical nucleons
in the same or different orbits. Then we will turn briefly to the case of two
nonidentical nucleons (proton and neutron) and the role of isospin and ex-
change terms in the residual interaction. In the next chapter we will consider
the case of larger multinucleon configurations of the form/1.
A complete treatment of many of these issues involves extensive and
sophisticated familiarity with the formalism of angular momentum and tensor
algebra. Indeed, many of the results can only be proved by rather formal
manipulation of the various angular momentum coefficients (3-/ symbols,
Racah coefficients, coefficients of fractional parentage, and so on.) and a deep
understanding of their symmetry properties. It is both contrary to the spirit of
our presentation and entirely beyond the competence of this author to present
this formal material fully. Moreover, there are numerous existing texts that
deal with it at great length in superb fashion. Two of the best are those of de
Shalit and Feshbach and de Shalit and Talmi, to which the reader is encour-
aged to refer. In the following presentation, we will make explicit use of their
results, but our principle effort is to provide a physical understanding and
motivation for the results and to present arguments for their plausibility. The
excellent recent book by Heyde also develops these ideas in a more formal
way. It is complementary to the present treatment and is highly recommended.
Before discussing the particular characteristics of multinucleon configura-
tions under the influence of residual interactions, it is crucial to discuss the role
67
68 Shell Model and Residual Interactions

of closed shells on the valence nucleons. As with the energy levels of single
nucleons in shell model orbits, it is only if the effects of closed shells can be
neglected that the study of multinucleon configurations will be applicable to
other than the lightest nuclei.
Since a closed shell is spherically symmetric, the interaction of a valence
nucleon in the state jm cannot depend on m: hence, the closed shell particles
have the same effect on all valence nucleons in a given; orbit. In multiparticle
configurations, the effects of the closed shell nucleons are independent of the
way (the relative orientations) that the individual jtn values are coupled to the
total J. Therefore, the closed shell can have no effect on energy differences of
these different J states. As discussed in Chapter 3, this does not rule out an
additive energy for the entire group of states and different; values can lead to
different shifts. Thus, the "rule" for spacings refers only to those within a given
configuration, not to the relative energies of different; shells in a major shell.
With these preliminaries in hand, we can now discuss two-particle configu-
rations. First, we need to determine which J values result from coupling of a
nucleon in orbit ;'j with one in orbit ;2. If jl *J2, these are simply the integer
values from l/j-;'J to I/, +; 2 |. If the orbits are equivalent, n,/^ = nj,jz
(which we shall often abbreviate to;'j =j2 when no confusion should arise) we
distinguish two cases— identical and nonidentical nucleons. For the latter
(proton-neutron) case, / takes on all integer values from 0 to 2j, as there is no
Pauli principle restriction on the occupation of identical m states. However,
for identical nucleons, one must explicitly consider the effects of the Pauli
principle, which requires the total wave function to be antisymmetric.
For identical nucleons, the isospin projection Tz = fz (1) + tz (2) = 1 and hence
the total isospin T = 1. (Clearly, T cannot be less than its projection, nor
greater than the sum of the individual tz values.) This is a symmetric wave
function. Hence, the space-spin part must be antisymmetric. When we impose
this requirement, only certain/ values are allowed. To see this, we have to look
at the particular m states occupied by the particles, 1 and 2. The Pauli principle
requires that

A properly antisymmetrized wave function that has this property is given by

where N is a normalization factor. The relation between the two Clebsch-

Gordon coefficients on the right-hand side is well known and given by the
phase factor (-l)2y~y. Hence

Since 2j is odd, this vanishes unless / is even: the only allowed J states for two
 The Shell Model: Two-particle Configurations 69

identical fermions in equivalent orbits are those with even total angular
momentum / = 0,2,4,...(2; -1). In Chapter 5, we will see that the same result
is even simpler to obtain in the m-scheme, which provides a general, though
sometimes tedious, way of finding the possible J states for any multiparticle
configuration.
With this result for identical nucleons in hand, we have an alternate way to
look at the complete set of J values 0 - 2j, available in the p-n system. We
consider the isospin structure of this system and use the result just obtained.
The p-n system can have T = 1/2 ± (-1/2) = 0 or 1 and hence T = 0 or 1. The
T=l case is identical in all respects to the preceding p-p and n-n T=l cases
and therefore consists of the even / values 0,2,4,...(2j -1). The T = 0 case is
antisymmetric in isospin so the space-spin part must be symmetric. By a
derivation exactly analogous to that just given, one finds that the T = 0 p-n
system contains the odd /values. Together, the T(p-n) = 0 + 1 parts give ally
values from 0 to 2/. (Note that, for j * jn, each two-particle J state is now a
mixture of T= 0 and T=l parts: a total nuclear wave function of good isospin
is constructed by coupling this two-particle state to the core.)

4.1 Residual Interactions: The ^-Function

We now turn to the energies E (jjyT) of a two-particle configuration of
identical nucleons denoted \jjjM), in the presence of a residual interaction.
We have noted earlier that for any central or scalar interaction, the wave
function can be separated into radial and angular parts. Since the radial
behavior of the two interacting particles does not differ for different./, many
results are independent of the detailed specification of the radial nature of the
interaction. This is certainly true for the ordering and relative spacings of
different J states: absolute values, of course, depend on detailed integrations
over the radial coordinates and on the strength of the interaction. Even in this
case, however, the results for different interactions are roughly similar.
Of course, with no residual interaction, all J states of the two-particle
configuration are degenerate, as shown on the left in Fig. 4.1. In the presence
of a residual interaction, the energy shifts (Fig. 4.1 right) relative to the
degenerate case are given schematically by

where the last step utilizes the Wigner-Eckart theorem. Most of the discussion
in this chapter centers on the crucial properties of these energy shifts due to the
residual interactions. Because of the independence of M, we shall generally
omit the magnetic quantum numbers in the discussion that follows. We have
already stated that it is possible to separate the radial and angular coordinates
for many residual interactions. To illustrate this, it is useful to consider the
simple 5-function interaction: by definition, this interaction vanishes unless
the particles occupy the same spatial position. The reason for choosing a 5-
function residual interaction is not simply mathematical convenience. More
importantly, it is preeminently a short-range interaction, and we know that the
70 Shell Model and Residual Interactions

Fig. 4.1. Schematic illustration of the energy shifts in a configuration with particle 1 in orbit j} and
particle 2 in orbit j coupled to various spins J.

nuclear torce, including residual interactions, has just this character. At least
qualitatively, a 5-function residual interaction reproduces many observed
properties of nuclei.
Moreover, it can be shown that a 5-interaction in a/" configuration is equiva-
lent to an odd tensor interaction. Such interactions are diagonal (see Chapter
5), in the so-called seniority scheme and as such are particularly useful for
treating multiparticle configurations, since many important results reduce to
the two-particle case. Thus, a discussion of the 5-interaction in | jjj} configu-
rations has profound implications throughout the study of nuclear structure.
We can write the 5-interaction as

where the second form expresses the interaction in polar coordinates. This is
 The Shell Model: Two-particle Configurations 71

particularly useful when separating the angular and radial parts. The l/r^
factor is necessary because the integration over the angular coordinates intro-
duces a factor 4;cr2.
Using the polar coordinate form and performing some straightforward but
tedious angular momentum algebra (see de Shalit and Feshbach, Chapter 5)
we obtain for the energy shifts in the identical particle configuration

where

and

FR depends only on the radial coordinates, while the quantity A results from an
integration over the angular coordinates.
Therefore, one obtains the extremely important result that the relative
splittings depend only on universal angular functions A, which are totally
independent of the nature of the central potential. They are also independent
of the principle quantum numbers nt and n2 (except for an overall scale
incorporated into the factors FR); for this reason, we shall not specify the n
values unless it is necessary in a particular case. Although Eqs. 4.6-4.7 apply
in detail only to a 5-function interaction, the general separability into radial
and angular parts is valid for any residual interaction that depends only on the
separation (r5 - r2) of the two nucleons.
We will soon discuss the meaning and implications of Eq. 4.5 at considerable
length. First, however, it is useful to give the expression for the specific case of
equivalent orbits /^ = ljr Hence /t + /2, must be even. Moreover, proper
normalization of the wave functions introduces a factor of 1/2 in the energies
that are now given by

where

and

Note that for / = 0, A <* (2j + 1)12, so that AE(fJ = 0) = const (2; +1)/2, that is,
72 Shell Model and Residual Interactions

the energy lowering of 0+ states is larger for large / and is in fact proportional
to the number of magnetic substates in the orbit/. This property is identical to
that defined for a pairing interaction. Here, its physical basis is that, for high
/, there are more magnetic substates spanning the same angular range of orbit
orientations. Hence, the wave function for a given substate is more localized
in angle. Two particles with the same I m I value thus have greater overlap and
hence a larger interaction. Similar overlap arguments will help us understand
many of the effects of residual interactions. They will be formalized later and
will pervade this chapter.
Returning to the general case of any jv j2, we note that, while the relative
energies of individual J values depend only on the angular structure of the
wave functions, the overall scale of the interaction and the average interaction
strength for particles in orbits jl and/2 depend on the radial integral (Eq. 4.6).
As we stated in Chapter 3, this will be largest for similar n^ and n2/2 values.
Shortly, we will calculate explicit numerical values of AEQJ^I), or rather, of
A (/v//)- First, however, it will help us understand two-particle configurations
and, later, multiparticle situations, if we momentarily ignore the analytic
formula and try to understand the basic results for the energies &E(j\j2J) from
simple physical arguments. It is remarkable how far this will take us.
We start with the obvious statement that the attractive 5-function interac-
tion can only be large when there is large spatial overlap between the orbits of
the particles. As we have seen, for a given ;t and;2, the overlap for different J
values depends on the orientation of the orbits in space. We shall see in Section
4.2 the explicit relation between the 3 -;' symbols in Eqs. 4.7 and 4.10 and the
relative angular orientation of the semiclassical orbit planes. For now, we
proceed more qualitatively. From overlap considerations alone, one might
think that the interaction would be largest either for J = 0 or J = Jrmax, which are
simply the two cases for which the angular momentum vectors j^ and ;2 are most
nearly antiparallel or parallel, respectively, and therefore those in which the
nucleons orbit the nucleus most nearly in the same plane.
While this simple view has an element of truth to it, the requirements of
antisymmetrization refine it considerably. Antisymmetrization, or the Pauli
principle, has enormous and profound implications throughout the study of
nuclear structure. We have seen how it determines the valley of stability,
validates the fundamental concept of independent particle motion in a dense
sea of nucleons, and gives the magic numbers and attendant shell structure.
These (or at least the first and last) are straightforward and obvious effects.
Others are subtle, even unexpected, and are certainly seldom appreciated.
Our case of two nucleons interacting via a short-range residual interaction is
just such a case. We start by considering the configuration of two identical
particles (two protons or two neutrons) in orbits jrjr
In this case, the two-particle wave function must be totally antisymmetric,
that is, antisymmetric in space, spin, and isospin. A careful understanding of
this requirement leads to some beautiful, remarkable, and profound results.
Since we are dealing with two identical particles, the states involved must
have total T=\ and hence T=l: they are symmetric with respect to the isospin
coordinates. Therefore they must be antisymmetric in space and spin
coordinates.
 The Shell Model: Two-particle Configurations 73

Thus far we have carried out the discussion in what is known as a// coupling
scheme, in which the total angular momentum jl of particle 1 is coupled to the
total angular momentum j2 of particle 2 to produce a final total angular
momentum /. One can also think of the problem in terms of the so-called LS
coupling scheme, in which the orbital angular momenta of the two particles lv
12 are first coupled to total L and the intrinsic spins (1/2 ft) ^ and s2 are coupled
to S = 1 or 0. Generally the jj coupling scheme is more useful in most nuclear
structure applications, but in the present case the L5 coupling scheme sepa-
rates out the angular and spin coordinates nicely and allows a simpler under-
standing of the effects of antisymmetrization.
To see how this works, let us take a particular example that gives a simple
result. Consider the case of two particles in a 1 d5/2 t,aJ} configuration as shown
in Fig. 4.2. The /values are / 1 = 2 and / 2 - 3, respectively, and in both orbits,
j = / +1/2. The allowed / values range from 1 to 6. If we picture the vector
coupling of/, and;2 to form various J values, it is clear that the orbital planes of
the two particles will overlap the most when J = 7mix and /mjn. We therefore
expect one or both of these to be the most affected (lowered) by an attractive
short-range interaction. To proceed further, we note that Jma is greater than
^ + /2 = 5. Therefore, the /mai state can only be formed by aligning the orbital
angular momenta to L = 5 and the intrinsic spin angular momenta to S = 1, and
then aligning L and S to / = 6.
Since 5=1, the two intrinsic spins point in the same direction and thus the
spin part of the wave function is symmetric. Therefore, antisymmetrization of
the total wave function requires that the spatial part be antisymmetric. Denot-
ing the angular part of the ith particle wave function by ^(r), the requirement
of antisymmetrization in the angular coordinates is equivalent to 0|10r1)0/20r2) =
-0>1(r2)^2(r1). However, the 5-function interaction is only effective when the
particles are in contact, when tl = rr At this point, this expression for the
antisymmetrization condition, however, requires that the wave function equal
its negative [0;1(r)^2(r) = -0^(00^(0], which of course can only happen if each
side vanishes. Therefore the wave function vanishes at the only point in space
where the 5-function interaction acts, so the residual interaction has no effect
whatsoever in this particular J = 6 state. Turning the argument around, a 5-
function interaction between identical nucleons can only affect states through
amplitudes in which S = 0, in which case the spin part of the wave function is
antisymmetric and the spatial part is symmetric and need not vanish at
I f j - rj =0. Interestingly, although the 5-function interaction has no explicit
spin dependence, its effects depend critically on the relative orientations of the
spins of the two nucleons. In other words, antisymmetrization introduces an
implicit spin dependence.
Clearly, to determine which states are affected by a 5-interaction we would
like to know which have 5 = 0. Unfortunately, except for the configuration
I sz,2/= 0), where 5 obviously must be zero, all two-particle configurations that
can have 5 = 0 amplitudes will also have 5=1 amplitudes. For example, it might
be thought that the / = 1 state of the I dsa L,a J} configuration would be pure
5 = 0 since this state is made by antialigning;, = 5/2 (/+1/2) and;2 = 7/2 (/ + 1/2).
Therefore the two intrinsic spins are also antialigned, giving 5 = 0. Though this
argument does identify the main component of the configuration, it is a bit
74 Shell Model and Residual Interactions

naive: one must be careful of mentally mixing the LS and jj schemes. In fact,
the/=! state can be made in three ways—by coupling/j and/ 2 to L = l.with
S = 0 as just stated, but also with the same L value and vector coupling S = 1 to
again give J - I and, finally, by coupling /5 and 12 to give L = 2 and then
antialigning a S = 1 vector to give 7=1.
In general, the relative amplitudes of S = 0 and 1 in a given / value for the
configuration, | (/jS^jj (/2s2);2;JM) are given by the 9-j symbol

Returning to the issue of determining which state is most affected by a <5-

interaction without actually carrying out the calculation of Eqs. 4.7 or 4.10,
there are three practical methods. All depend on the fact that the 7min and Jma
states are most nearly coplanar, and therefore one of these will be the most
lowered, while the other is unaffected. Determining either, then, answers the
question.
The most straightforward approach is the one we have just used in the
I d^f^T) case, namely to look for a state that is pure 5=1: this state will be
unaffected, as we have seen.
There will be such a state (either /^ or Jmm) with pure 5=1 only when there
is an allowed / value that is greater than (/} + /2) or less than (l^ - /2). This oc-
curs, for example, for J = 6 in | d5/2 fin J), or J = 1 in | d5/2 g7/2 J) where Lmin = 2
(see Fig. 4.2), for J = 8 in | d5/2 h11/2/), but not for any state in | g^ hn/2 /} or for
the allowed (even) J values in any configuration of the type I fJ). A useful rule
is that a state with pure 5 = 1 always exists if jl *j2 and/ lower is / + 1/2.
When no such state exists, a second method uses the fact that a 5-interaction
only affects half of the states in a multiple!; either those with even / or those
with odd /. We have seen an example of this in the I ds/2f7/27) configuration
where the /* = 6~ level is unaffected, and the 1- level is lowered the most. In
contrast, in the I ds/2 g7/2 J} configuration, /min again equals 1. But here
^rnin = I A ~ U = ^ and therefore the 1+ state can only be found by antialigning
5 = 1 to L = 2. Thus, here it is the 1+ state that is unaffected, while the 6+ state
is lowered the most. These two cases are just examples of the general rule that
for positive parity configurations only even / levels are lowered, while for
negative parity only odd J levels are lowered. This rule is clear from the
restrictions on the right in Eq. 4.7: for positive parity l^ + /2 is even, so /: + /2 - /
is even only for / even and vice versa for negative parity. We use this rule,
which arises simply from the parity of the interaction matrix elements that
involve the spherical harmonics YLm(0) where 9 is the angle between the
orbital planes (see Section 4.2), as follows. For anyjJ2, one of Jrnin or Jmai will
be odd and the other even. That state that falls into the class of states affected
by the interaction will be lowered the most, and the others in the same class
successively less so as their J values deviate more from J max or Jmm. . Let us
J

consider a couple of examples. Starting with the familiar I d5/2f7/27) case, ymln
and 7mM are 1 and 6, the parity is negative and thus only the odd / states will be
 The Shell Model: Two-particle Configurations 75

Fig. 4.2. Semi-classical illustration of the coupling of intrinsic spins and orbital angular momenta
for two configurations (d$/2 f?/2) and (dJ/2 g?/2)

affected, that is J* = 1% 3~, 5~. Therefore the/= Instate is lowered the most, the
J = 3~, 5" states successively less so, and the / = 2", 4~ and 6 states not at all.
For the d5/2g7/2/} configuration, /min = 1, /max = 6, as before but here n = + so
that / = 6 is lowered the most and the order is (from lowest to highest) 6\ 4', 2+
(1% 3+, 5+ degenerate at the unperturbed position). Finally, for |g7/2 hlu2/},
/mjn = 2, /max = 9, TT = -, and so the order (again, lowest to highest) is 9~, 7 , 5 , 3
(with 2', 4-, 6 ,8~ unperturbed). We note an interesting point, namely that, just
as with S values, although the 5-function has no explicit parity dependence, the
resulting energies of states \JJ2J) are, in fact, different for the two different
parities.
Although no states (other than | s21/27 = 0)) have pure 5 = 0, a third method,
which is formally incorrect, emphasizes they; coupling picture and does always
give the right answer. To illustrate this procedure, in I g7/2hn/27), all states J = 2
to 9 can be formed with S = 0 or 1. The naive argument would say that, since
y\ = /t - 1/2 and y'2 = 12 + 1/2, the 7min = 2 state formed by antialigning j^ and y'2
would correspond to S = I and therefore be unaffected by a 5-interaction. The
same argumentation would imply that the ,/max = 9 state is S = 0, and therefore
would be lowered the most.
76 Shell Model and Residual Interactions

This approach can be used for equivalent orbit cases to obtain a very
important result. Consider a configuration 1/7) such as |(f7/2)27). The/value
corresponds to / + 1/2, and therefore the antialigned /mjn state, 7 = 0, should
have 5 = 0 and be most affected. Again, although the argument is not really
correct, the result is. Since the overlaps of the particle orbits are reduced with
increasing J, the excitation energies increase monotonically with /. Note that
the same result applies for any case of identical particles in equivalent orbits:
the J* = 0+ state will always lie lowest. Note also that, in this respect, the 5-
function interaction resembles the well-known effects of the pairing force that
is designed (defined, actually) to lower states in which pairs of identical
particles are coupled to spin 0. As pointed out just after Eq. 4.10, and as seen
in the upper panels in Fig. 4.3, this effect of the 5-interaction in \fj) configu-
ration is greater for higher; (proportional to (2/ + 1)).
This lowering of the 0+ state is an extremely important result. Ultimately, it
is the underlying reason why all even-even nuclei have 0+ ground states and,
often, large spacings to the 2+: level. It directly explains this result only for
nuclei two nucleons away from closed shells. However, we shall see in Chapter
5 that our two-particle result can be generalized to f configurations, and by
extension, to wave functions that are linear combinations of several/" configu-
rations. Further generalizations to vibrational excitations and deformed
nuclei will be seen in later chapters.
There is an easy geometrical way of viewing this case of identical nucleons
in equivalent orbits that provides a physical rationale for the lowering of the 0+
states. Intuitively, it might seem that the overlap of the two particles in the /min
and /max states would be comparable: the two orbits are essentially coplanar in
both cases. Once again, however, the Pauli principle plays a key role in
distinguishing these situations. We show a schematic illustration in Fig. 4.4. In
the / = /max state, the near alignment of the two / values implies nearly coplanar
orbits in which the two particles orbit in the same direction. The Pauli
principle, however, forbids contact. In effect, this means that the two particles
must repel each other at short distances: therefore they track each other
around the nucleus on opposite sides of an orbit so that they always remain
apart. Thus, there can be no 5-function interaction between them. For Jn = 0+,
the orbits once again are coplanar but now the two particles orbit in opposite
senses. As they do so, their separation will vary but the average separation will
clearly be much less than in the /maj state, "contact" situations occur, and a
large 5-function interaction results. The actual values oiAfj^f) for a number
of different spin combinations are summarized in Table 4.1, and Fig. 4.3 shows
several examples (including those used most often in the preceding discussion)
of AEdjyf) values under the influence of a 5-function interaction. In studying
Table 4.1, recall that the interactions are attractive so that larger values of
A(jJ2J) correspond to lower-lying levels. Perusal of the table and figure shows
that the preceding rules are always satisfied.
It is possible to summarize these results succintly, as is done in Table 4.2.
(The trigonometric functions in column 6 of Table 4.2 will be explained in
Section 4.2.) To illustrate the construction of the table, let us consider the top
row. A little thought, or working out a few examples of the preceding rules,
 MULTIPLE! SPLITTINGS; 8 INTERACTION ( I d e n t i c a l P a r t i c l e s )

Fig. 4.3. Energy shifts for a 5-function residual interaction for identical nucleons in several different orbit combinations.
78 Shell Model and Residual Interactions

IDENTICAL NUCLEONS
EQUIVALENT ORBITS
Fig. 4.4. Pictorial illustration of the motions of two identical nucleons in equivalent orbits for the
cases of maximum and minimum total spin J, showing the effects of the Pauli principle.

will convince the reader that, for positive parity, the state most affected will be
•'rain tf 0\ ~/ 2 ) ig even (as m Idszgjra^)). Note that even values of (jl -/2) are
equivalent to odd values of (jl + j2). Moreover, since only even/states are
affected for positive parity states, odd values of (j\ + ;2) are the same as odd
values of (jl +j2 + J). The consistent element in the table is that, if yt +J2+J is
odd (even) for the affected states, then the lowest-lying state will be
•Anm = U ~ h I (•'ma* = A + A)- We wil1 encounter the important role of the
quantity^ + /2 + J later in dealing with interactions in p-n multiplets.
To illustrate how these ideas relate to empirical data on nuclei with two
valence nucleons, we recall Fig. 2.5, which showed the yrast states for the
typical shell model nuclei 210Po, 210Pb, and 134Te. We repeat those data here in
Fig. 4.5, along with the predictions for a 5-interaction normalized to the
Ei\ energy for the relevant configuration [(h9/2)2 for 210Po, (g9/2)2 for 210Pb
(these two cases are actually identical as seen in Eq. 4.10), and (f7/2)2 for 134Te.]
The agreement is quite good. Note the strong lowering of the 0+ ground state.
From data such as this, known for many nuclei, it is possible to estimate the
absolute strength of the interaction. One obtains, as a rough guideline, that
AE (j\ j2 J) ~ (30 MeV/A) A(Jl J2 J), where the A in the denominator is the
Table 4.1. Relative J state energy values for various identical two-particle configurations I / j J) with an attractive 5-interaction.*
J
/, A WX «f), n 0 1 2 3 4 5 6 7 8 9 10 11

d5* dM
5/2 5/2 f» + 3.00 0.685 0.286
dM ^
f» 0 0.173 0 0.457 0 1.29
S?<2 %nei
7/2 7/2 *» f» + 4.00 0.95 0.467 0.233
&7/Z f» 0 0.127 0 0.312 0 0.599 0 1.52
»»,» h1M
11/2 11/2 '»« 'iw + 6.00 1.47 0.785 0.493 0.318 0.180
*,» W 0 0.084 0 0.196 0 0.329 0 0.509 0 0.818 0 1.89
ds«
3/2 5/2 ^
P3« «H
+ 0 0.343 0 1.14
d
3* fM 2.4 0 0.686 0
PM dsa
d» 87/2
5/2 7/2 f
« + 0 0.230 0 0.518 0 1.40
d5a ^
f» 3.41 0 1.14 0 0.518 0
<» &7/2

d5* SM
5/2 9/2 f
« h» + 2.85 0 1.04 0 0.490 0
d5, hM 0 0.258 0 0.558 0 1.47
f» 8s/2
d
s. in*
5/2 11/2 h
iifl + 0 0.279 0 0.587 0 1.52
d^sa \m 2.65 0 0.979 0 0.470 0
f« \m

"The table gives values olAyjfl from Eqs. 4.7 and 4.10. Dashes indicate J values not allowed for the given configuration. States that are unaffected by the interaction (i.e.,7 even for
TT = -, J odd for x = + configurations) are given a value of 0. Other numbers are proportional to the decrease in energy of the state /.
80 Shell Model and Residual Interactions

Table 4.2. Rules for the effects of a 5-function interaction on two-particle identical nucleon
configurations
Configuration Semi-
Category States classical
Parity Affected Lowest State dependence
A+/2 h+h + J one*
Positive odd odd even/ mill
tan 6/2
Negative even odd odd/ (antialigned)
Postive even even even/ '-. cot 6/2
Negative odd even odd/ (aligned)

*6 is defined by

mass number. Recalling that AO"2 7 = 0) = (2; +1)12, this gives a typical lowering
of the ground state of several MeV in light nuclei and of 1 to 1.5 MeV in heavy
nuclei where the interaction strength is only ~ 200 ke V. Both of these are well-
known features of the data (e.g., Figs. 2.6-2.10), and again show that a 5-
function naturally produces the famous "energy gap" in even-even nuclei
usually associated with the pairing force (see the following).
In all these examples, there is another important feature we have not yet
commented on: the lowest level for a given multiple! is substantially lowered,
but the differences in interaction strength for the others monotonically de-

Fig. 4.5. Comparison of experimental and calculated low-lying even spin (yrast) states in three
nuclei with two valence nucleons. The orbits used for the two identical nucleons are indicated in
each case.
 The Shell Model: Two-particle Configurations 81

crease. That is, there is a relative compression of levels near the unperturbed
position. This is not just an accident of the 3 - / symbols, but has a simple
physical origin that we shall discuss shortly. As we noted in Fig. 2.5 and, here
in Fig. 4.5 for the case of jl = 72> it is also a well-known empirical effect
characteristic of the low levels of many "shell model" nuclei.
We have been discussing the effects of a 5-interaction between identical
nucleons. Such states have iz (1) = tz (2), hence Tf = ±1, and T = 1. The
proton-neutron system also exists in a T= 1 state. By charge independence of
the non-Coulomb part of the nuclear force, the p-n T= 1 system must then also
satisfy Eqs. 4.5-4.7. Indeed, the familiar statement of charge independence
that p-p, n-n, and n-p forces are equal applies specifically (and only) to the
T = 1 mode for the p-n system.
As we have seen, however, the p-n system can also exist in a T= 0 state for
which there is no need for equality to the p-p or n-n forces. Empirically, in
fact, the T- 0 interaction seems to be significantly stronger than the T= 1 (see
the following). This T= 0 coupling is extremely important in nuclear structure,
as it is now thought to be responsible for single-particle configuration mixing
and the onset of collectivity, phase transitions, and deformations. We shall
return to these points in later chapters. For now, we are interested in simple
two-particle p-n configurations in shell model (noncollective, nondeformed)
nuclei under the action of a 5-interaction.
In order to address this issue, we must deal with a specific complication thai
arises in the p-n system. Suppose we imagine such a system occupying levels
a and b as shown on the left in Fig. 4.6a. Then, if we treat the proton and
neutron as two states of the same particle (the nucleon), the orbits or the
charges call be exchanged indistinguishably. Thus, the wavefunction for the

Fig. 4.6. "Direct" and "exchange" configurations for protons and neutrons treated as indistin-
guishable particles (a) filling the same shell, and (b) where the neutron shell corresponding to the
valence protons is already filled.
82 Shell Model and Residual Interactions

two-particle system will have components of all the types illustrated. An

interaction matrix element will then contain, for example,

The four terms on the right correspond to the four cases of particles 1 and 2,
each occupying levels a or b, and each identified as either a proton or a
neutron. The term in a in Eq. 4.11 is a direct term that should be relatively
large since the overlap of the wave functions on the two sides is unity. The last
term keeps the particles in the same orbits (1 in a, 2 in b) as on the left side but
exchanges their type (p -» n, n -» p). This can also be large if the interaction
contains terms that can change a proton into a neutron and vice versa. The
second and third terms involve overlaps in which the particles change orbits:
therefore they are normally small. In principle, however, all must be taken into
account. Clearly, this is both complicated and tedious. The isospin formalism
for doing so is well known and is discussed in standard texts, so we will not
consider this situation. Here we are largely concerned with medium and heavy
nuclei, which greatly simplifies the problem since protons and neutrons are
usually filling different major shells. This situation is illustrated in Fig. 4.6b,
where the neutron shell corresponding to the proton shell is already filled:
hence, exchange matrix elements such as those in the 7 and <5 terms in Eq. 4.1 1
(the rightmost two terms in Fig. 4.6a) are impossible ("blocked"), since the
neutron orbits in that shell are filled. The only remaining exchange term is one
of those expected to be small. We can consider only the direct term to good
approximation in such a case. This is equivalent to treating the protons and
neutrons as distinguishable particles. Note that, while we can choose to do so
in this case, because of the blocking of exchange terms, it is not mandatory: we
could (and will momentarily) use the isospin formalism as well. The
proton-neutron and isospin formalisms are equivalent ways of obtaining the
same result. Which one is used in a given case is a choice based on practical
simplicity. When exchange effects are known, a priori, to be inconsequential,
the proton-neutron formalism may be simpler.
In the present case, treating the protons and neutrons as distinguishable
means that the Pauli principle places no restrictions on their coupling. Their
wave functions need not be antisymmetrized. We can write the two-particle
p-n wave function as y/^ = 0 0n. To see the relation to the isospin approach,
this can be rewritten as \y n = l/2[(0 <t>n + 0n0 ) + (0 0n - 0n0 )] • The first term is
symmetric with respect to interchange of protons and neutrons, the second
antisymmetric. These two terms therefore correspond to T = 1 and T = 0,
respectively, and the energy shifts in the presence of a residual interaction can
be written as the average of their values in the two isospin channels, that is, as
1/2[AE(T = 1) + A£(T = 0)]. In the like-nucleon case, antisymmetrization
requires that the S = 1 (symmetric) spin coupling be accompanied by an
antisymmetric spatial part of the wave function (which vanishes when the two
particles are in contact): such states do not feel the 5-interaction. Here, there
is no such limitation. Both 5 = 0 and 5 = 1 states can be accompanied by
 The Shell Model: Two-particle Configurations 83

symmetric spatial wave functions and will be shifted by the interaction. Gen-
erally, the interaction need not be of the same strength in the two cases, and so
we specify two interaction strengths Vs_0 and Vs=l

where it is implicit that the Vs=t> (VJ=1) term acts only on the S - 0 (5 = 1)
components.
We can get the same result in terms of isospin. We saw before that the
unsymmetrized two-particle wave function can be written in terms of an
average over T = 1 and T = 0 parts. In this case, the wave function for each
isospin term must be separately antisymmetric. Specifying to a 5-function
interaction, the only states affected must have symmetric spatial wave func-
tions. Therefore, the isospin-spin part must be antisymmetric: hence, the
T = 0 part goes with 5=1 and the T= 1 with 5 = 0. Again, we get S = 0 and 1
terms as in Eq. 4.12. (Recall that the total nuclear wave function must have
good isospin, which is obtained by coupling the isospin of the two-particle
system to that of the (T * 0) core.)
Thus, the analogues of Eqs. 4.5-4.7 for a p-n system under the action of a 5-
interaction given by Eq. 4.12 become:

where

and

The 3 -;' symbol is identical to that appearing in the like-nucleon case, but now
there are two terms with different /,/ ,jn dependencies. Note that, in the first,
or Vs=0, term in Eq.4.15, only half of the levels are affected, namely those with
even / for positive parity or with odd J for negative parity. This is the same con-
dition we saw for like nucleons, as it must be since this is the 5 = 0 (T- 1) term.
The other/values are then affected by the first of the Vs^ terms. For; *jn, the
second Vs=l term itself affects all /values, and small / values the most.
As in the like-nucleon case, this equation simplifies for equivalent orbits
84 Shell Model and Residual Interactions

Here we can easily see the explicit relation to the isospin formalism. The
T - 1 p-n interaction must be identical to the p-p and n-n interactions. We
saw earlier that, for equivalent orbits, only even / states are allowed for the
T = 1 p-p and n-n systems, and hence, for a p-n system, which has all / values
from 0 to 2;', the even/values must have T= 1 and the remaining levels, namely
those with odd /, must be T = 0. Thus, in Eq. 4.16, the first term corresponds
to the interaction in the T= 1 channel and the second term to the T=0 channel.
We now need to consider the relative strengths Vfc0 and V5=1. From the fact
that the deuteron has an S = 1 ground state, it is clear that V5=1 is stronger than
Vs=0. However, there is additional evidence for this from such simple data as
neutron separation energies that is directly applicable to nuclei with all A
values. As we have just seen (Eq. 4.16) the T = 1 and 0 interactions can be
associated with the S - 0 and 1 terms, respectively. We saw in Chapter 1 from
the separation energy data that the nonpairing, like-nucleon (T- 1) residual
interaction is, on average, repulsive, where by the phrase "on average" we
mean averaged over all final / states and by "nonpairing" we mean excluding
the 0+ state (if any). So, by charge independence, the p-n T = 1 interaction
must on average also be repulsive. Yet, we also noted in Chapter 1 that both
S(p) and S(n) increase with increasing numbers of particles of the opposite
type. The interaction between protons and neutrons has both T = 0 and 7" = 1
components. So, on balance, the total (T- 0 + T= 1) p-n interaction must be
attractive. This can only occur if the T = 0 component is both attractive and
stronger than the T= 1, that is, if the \Vs=l\ is greater than Vs=0 .
Of course, the strength of the two isospin components of the interaction can
also be obtained by fitting actual p-n multiplets (groups of states with pure
proton and neutron configurations; and jn and /values ranging from \j -jn\
to jf + ;J. Schiffer and True and Molinari and co-workers have carried out
extensive surveys of this type near all closed shells from 160 to 208Pb. We will
discuss their results in Section 4.2 in terms of a simple geometrical analysis.
Here it is useful to convey a feeling as to how the data on individual isospins
can be deduced. The nuclei near 208Pb offer a nice example. Consider, for
example, the states of ^Po^ in a Ih9/2li13/z7) two-proton T = 1 multiple!.
(These can be found from the 209Bi (3He, d) 210Po reaction since 209Bi has a single
proton in the lh9/2 orbit.) The energy shifts in this multiplet can be used to
extract the (lh9/2 iiiy2) T= 1 interaction. The same multiplet exists in20883BiU5
as a particle-hole p-n multiplet. The energy shifts AEQJ-JJ) can be converted
(see end of chapter) to an equivalent set of particle-particle shifts and the total
p-n interaction obtained for each/state. The difference of the T= 1 and total
interactions then yields the net T = 0 strengths. Extraction of T = I and T = 0
strengths is even simpler in the case of equivalent orbits (/ = y'J, of course,
where the even and odd J states directly give the T = 1 and 0 interactions,
respectively. This approach is useful in light nuclei where the protons and
neutrons are filling identical orbits (e.g., the f7/2 orbit in 42Sc).
To illustrate the application of these ideas, we consider the classic example
of 1738C121. Since the N = 8 to 20 neutron shell is filled, this is an appropriate
case to ignore exchange terms. In the lowest-lying states, the configuration is
(d.,^ f7/2/j) giving states J = 2 ,3", 4-, and 5 . Since Vs^ > Vs=0, the second group
 The Shell Model: Two-particle Configurations 85

Fig. 4.7. Comparison of low-lying empirical and calculated energies for 38C1. The two panels on
the right correspond to calculations with a two-body 5-function residual interaction, assuming two
different orbits for the proton. Clearly, the (d3/. f 7 _) configuration is favored. The calculation on
the left uses the empirical levels of the (d.^"1 f?/2) particle-hole configuration in 40K in conjunction
with Eq. 4.34 to predict the particle-particle levels of 38C1. (See deShalit, 1974.)

of terms in Eq. 4.15 will generally dominate and the overall ordering of levels
in the p-n system will tend to be contrary to that in the like nucleon case.
Moreover, whereas only half the states are affected for like nucleons (/ odd for
n = -\J even for n = +), all states will be shifted in the p-n case. We therefore
may expect the lowest level to be the even J state with highest overlap, the
/ = 2~ level. The 38C1 experimental spectrum and that calculated with
Vs=1 = 2F5=0 are shown on the right in Fig. 4.7. (The part on the left describes
an alternate approach to calculating 38C1, to be discussed near the end of this
chapter.) The 2-level does in fact occur lowest, and the agreement is reason-
able. The figure also shows that the calculated levels for an alternate configu-
ration with the same / values, (p3/2p_ f7/2J, have a rather different pattern since,
here, the orbital phase factors in Eq. 4.15 are different (/ + ln is now even) and
the J = 5, 3 set is lowered relative to the J = 2,4 pair in disagreement with the
data. This indicates how one can even sometimes suggest; configurations and
/"values by examining energy sequences and spacings in p-n multiplets.

4.2 Geometrical Interpretation

Having dealt extensively now with both like and unlike two-particle configura-
tions under the influence of a 5-function interaction, we have gained a feeling
for the physics behind the analytic results that can be obtained. The physics
revolves around the overlaps of the two-particle wave functions. It is possible
to approach this entire subject from an alternate viewpoint and actually derive
the typical behavior of the 3-y symbol in Eqs. 4.7,4.15, and 4.16 from a simple
geometrical analysis, which will give us additional insight into the interactions
in two-particle configurations.
We commented earlier that the characteristic and typical behavior of that
3 - j symbol is a gradual reduction in the spacings as the excitation energy
increases (as the interaction weakens). This is not one of those annoying
T

Fig. 4.8. Definition and schematic illustration of some of the ideas used in the geometrical analysis
of short-range residual interactions.

"accidental" effects of Clebsch-Gordon coefficients that plague many stu-

dents, but rather it has a very simple physical origin. In pursuing this we will
better understand why this 3 -/symbol behaves as it does. Moreover, we will
see that the energies &E(jj2I) exhibit the same basic pattern for any jJ2 and
that this pattern simply reflects the spatial overlaps of the particles and Pauli
principle effects.
We start with the semiclassical concept of the angle, 9, between the angular
momentum vectors j^ and ^ (hence between the orbital planes) of the two
particles as illustrated schematically in Fig. 4.8. Then

or

From here on, for simplicity, we take the case of identical particles in
equivalent orbits (j\ = j2 -./) and assume that/, 7 » 1 so that terms like./(./ + 1)
can be approximated by J1. Then,

Note that 0=0° corresponds to high J and 6 -180° corresponds to low J. Thus,
for ;; = jv G = 180° corresponds to J = 0 and 0^ -» 0° to J = 7mai = 2; - 1.
Before proceeding, we first make use of some simple trigonometric equa-
tions. From sin2 0=1 - cos2 6, we obtain
 The Shell Model: Two-particle Configurations 87

And, from sin we get

We also note that tan Now, the symbol in Eq. 4.10

can be written

A good approximation to this for large ;', J is

Neglecting quantities of the order of unity compared to /, J we get

Hence,

Using the relations for sin 912 and tan 012, we have

or, finally

This extremely simple result expresses the shifts in different / states for a 5-
interaction between two identical particles in equivalent orbits. It was derived
for large j, J, but is remarkably accurate even for low spins (e.g., as low as
j = 3/2 and 7 = 1 but specifically not for J = 0). The function tan 9/2 is plotted
against 9 in Fig. 4.9. Since tan 9/2 ~ 9/2 for small 9 and goes to infinity for
88 Shell Model and Residual Interactions

EQUIVALENT ORBITS

Fig. 4.9. Dependence of the <5-f unction residual interaction strength (lower values correspond lo
mDreattractiveresidualinteracu'ons)fortwoparticlcsinequivalentorbits. (Left)The7"=l (./even)states. (Right)
The 7 = 0 (/odd) states. The analytic expressions are indicated above their respective plots.

8 -> 180°, Eq. 4.18 simply states that the energy shifts become large (and
negative since the force is attractive) for 9 ~ 180° (for small J where the two
angular momenta are antialigned), while the smallest effect occurs when
J = 7max = y + y, since 0= 0°. Moreover, the curve tan 0/2 becomes asymptoti-
cally flat for large J, giving a geometrical interpretation to the compression in
spacings discussed above and illustrated for 210Po, 210Pb, and 134Te in Fig. 2.5.
Note that this formula automatically reflects the Pauli principle arguments
discussed earlier, in which the 5-interaction affects only 5-0 states. The Pauli
principle appears here through the 3-y symbol and, in particular, through the
spin angular momenta of ±1/2 appearing in it.
As we have noted, identical nucleons in equivalent orbits have T= 1, so Eq.
4.18 applies to them. For proton-neutron configurations we have to consider
both 7'= 1 and T- 0 parts. We still restrict ourselves, though, to equivalent
orbits. A T= 1 p-n system, which cannot be distinguished from the T- I p-p
and n-n systems, consists of even / states and is described by Eq. 4.18 as well.
However, in a T = 0 p-n system (odd ./ states ) AE n (fj) has a different./
dependence, reflecting the different./ behavior of the second (V^) term in
Eqs. 4.15 and 4.16 compared to the first term.
For the odd 7, T = 0 case, a similar analysis gives (again for;' = y'J
 The Shell Model: Two-particle Configurations 89

The behavior otAE(fJ) for T= 0 is also indicated in Fig. 4.9. We recall that
this only applies to equivalent orbits / = jn. While the T = 1 interaction is
largest (lowest-lying on the plot) for 7min (9 ~ 180°) and is smallest for
/ = /mM (9-> 0°), the T = 0 expression is large for both /min and 7mM. For T = 0
and 9~1800,&E(fJmit)~l/cos9--*°<>-,foT9=Q0,AE(j2JmJ~col9-^<>°. Both the
r = 0 and T= 1 expressions are small for 0= 90°.
All these features can be easily understood physically. The interaction
should be small for 9 = 90° for both T = 0 and T = 1, since the particles are
orbiting in nearly perpendicular planes and are seldom close enough to inter-
act. For T - 1 (which, by charge independence, means we can re-use the
identical-particle arguments), the interaction is strong when the two nucleons
orbit in opposite directions (J = 0,9= 180°). However, it vanishes when they
orbit in the same direction (Jmsa, 9 = 0°) since, then, the two particles have
identical quantum numbers and the spatial wave function is required to be
antisymmetric: it must vanish if the nucleons "touch." The Pauli principle
effectively introduces a short-range repulsion. The only way the particles can
orbit in the same direction and yet not touch is if they circulate out of phase at
opposite ends of an orbit diameter. This gives an interaction that is small for
small 9, but large for large angles in agreement with Fig. 4.9. The basic idea is
the same as for the identical particle T= 1 case (Fig. 4.4). For the T= 0 case we
treat the particles as distinct and, for both the small and large / extremes, the
orbits are nearly coplanar. Since we need not worry about antisymmetry, there
is no restriction on phasing, and "contact" is abundant, leading to a strong
interaction for both 0- 0° and 0= 180°.
Empirically, these effects are well documented as shown by the examples in
Fig. 4.10 taken from the aforementioned empirical analyses of p-n multiplets
throughout the periodic table by Schiffer and True. Note the interesting point
that for even J, T = 1, the empirical interaction is actually slightly positive
(repulsive) for small 9 (high /). A 5-function interaction cannot give this: at
best, it vanishes near J - Jmm. Such an analysis clearly shows the need for a
separate repulsive component in the residual interaction. Several studies have
successfully carried out multipole analyses of these effective residual interac-
tion, incorporating dipole, quadrupole, etc. components. Evidence for a
sizable quadrupole component has been found. This multipole varies as
P2(cos0) where, again, 0is the angle between the two orbits. As is well known,
this function crosses zero at 9 ~ 55° so that even for an overall attractive quad-
rupole term, the interaction is actually repulsive for angles between 55° and
125°. This is just the region where Fig. 4.10 shows positive (repulsive) empirical
T = 1 interactions. This repulsive aspect should not be surprising. We have
already encountered it. We noted in our discussion of separation energies in
Chapter 1 that the like nucleon (T = 1) nonpairing residual interaction was, on
90 Shell Model and Residual Interactions

balance, repulsive (S(n) decreases with increasing N). From this empirical
fact, we also deduced that the T-Q interaction is on balance stronger (more
negative) than the T=l. This is also evident in Fig. 4.10.
Finally, note that for J* = 0+, the interaction deviates from the geometric
expression. The 0+ behavior, however, is physically reasonable. As with the
like-nucleon case, the interactions are ordered by;': they are largest for large/.
The larger the) value, the more magnetic substates there are, and the smaller
the permissible angular range of an orbit for a given m. Thus the orbit planes
are more tightly defined and the overlaps of particles in ±m substates are
greater.

Fig. 4.10. F.mpirical proton-neutron multiplets for two particle equivalent orbit configurations for
comparison with the behavior shown in Fig. 4.9. The curves arc drawn through the data (Schiffer,
1971).
 The Shell Model: Two-particle Configurations 91

Thus far, for simplicity, we have carried out the geometrical analysis for the
simple case of/, =/2. For/, */2 and the identical particles, we saw in Section 4.1
that either the 7^(0^ 180°) or the/ max (0^ 0°) state can be lowest depending
on the particular/values and their/ = / ± 1/2 character. Table 4.2 summarized
the different cases leading to these two situations. These two categories of
two-particle configurations should be and are reflected in the geometrical
analysis. One obtains two curves now, of which one is identical to Eq. 4.18
(AE ~ tan#/2) giving the lowest energy for the antialignment of the two values
(7^, 9 close to 180°), and the other curve goes as cot0/2 so that the lowest
energy occurs for parallel alignment (7max) and 9 close to 0°. The correspon-
dence of these two trigonometrical functions and different sets of/ : ,/ 2 values is
made explicit in the sixth column of Table 4.2. Finally, note that the equiva-
lent-orbit situation is actually a special case of this. Here, n = +,/, + /2 = 2j is
always odd and so the land/2 dependence applies and the ./min (in this case ()*)
state is lowest.

Fig. 4.11. Comparison of empirical and calculated multiple! splittings for two-particle configura-
tions of noncquivalent orbits (Sehiffcr, 1971).
92 Shell Model and Residual Interactions

Fig. 4.12. A geometrical analysis of the lh<J/2g9/2/> p-n multiple! in 210Bi. The empirical levels are
shown on the left along with thesemiclassical angle between the orbits of the two nucleons. The
right side shows that the levels split into two families, according to7 even or J odd. The solid lines
are drawn to connect the points.

For the p-n case, a similar analysis again leads to two distinct curves as in
the T= I and T- 0 cases for equivalent orbits. However, the classification is
slightly different. Recall that, for equivalent orbits, the T= I states are even,/
only. Thus /, +j2 + J-2j + J is odd. The T = 0 case, with J odd, has /, + j2 +./ =
2; + J even. It is this distinction that persists when;^ *jz for a p-n multiple!.
Again, we obtain two curves, but distinguished according to the odd or even
character o f y , + y'2 + J and describable by geometrical functions of 9 very
similar to Eqs. 4.18 and 4.19. This is beautifully illustrated by the data for
several multiplets collected in Fig. 4.11, and exemplified in depth for 210Bi in
Fig. 4.12. In all these cases the empirical energy distributions within the p-n
multiplets follow the expected energy patterns quite well.
One last point worth mentioning is that extensive surveys of empirical p-n
interaction multiplets show that the strength of the interaction, especially in
T = 0 states, smoothly decreases with increasing mass. This is quite plausible
since the average radius of shell model orbits increases with higher oscillator
numbers, while the interaction range is constant so that the average interaction
strength decreases. In heavy nuclei, typical V n interaction matrix elements
are ~ 200 to 300 keV but of course this depends on the orbits involved.
What is perhaps most important to emphasize in concluding this part of the
discussion is that, without ever having dealt with the radial parts of the wave
functions, or indeed, calculating anything, it has been possible to predict the
qualitative energy ordering of the different./ states in two-particle configura-
tions. Moreover, exact quantitative results for the relative spacings involve
only the evaluation of a single 3-/symbol. (Of course, the absolute spacings
 The Shell Model: Two-particle Configurations 93

depend on the radial integrations and the strength of the interaction.) This is
but one example of how far one can go in a shell model treatment of multipar-
ticle configurations by invoking only very general arguments.

4.3 Pairing Interaction

We re-emphasize that these results for the 5-function interaction are of more
than passing interest since, representing the short-range interaction par excel-
lence, this interaction simulates in many respects results from more realistic
short-range interactions. In particular, it is closely allied to the pairing interac-
tion specifically designed to mock up a strong, attractive interaction in the / = 0
configuration of two identical nucleons. The motivation is similar to that for
the 5-function interaction—the pairing interaction is only effective when the
particles have extremely high spatial overlaps. Formally, one can define the
pairing interaction by

where G gives the overall strength in the interaction. Note that this interaction
is attractive and, by definition, only effective for 0+ states of identical nucleons
in equivalent orbits. It is not, however, limited to diagonal matrix elements
O^O* I Vpair I/Y^X but rather allows nondiagonal scatterings, </'t20+1V alr l/320+),
in which the pair of particles switches to another orbit as a pair. This feature is
critical to the build-up of pairing correlations and the so-called pairing gap in
even-even nuclei, and will be treated in more detail in a later chapter. For
diagonal matrix elements I/, = j2 - j3 = j4) the pairing interaction strongly
lowers the 0+ state without affecting the others.
Both the 5-function force and the pairing force are intended to represent
the short-range component of the nuclear interaction. However, the residual
interaction also contains a long-range component, that, as we shall see, is

Fig. 4.13. Comparison of levels of a (/ = 7/2)2 configuration for a 5-function and a pairing
interaction.
94 Shell Model and Residual Interactions

crucial in producing collective properties and nonspherical nuclei. It is com-

mon to mock this up by the so-called quadrupole interaction. (In recent years
higher order multipoles, such as hexadecapole interactions, have also been
considered, but we shall ignore their effects here.) The combination of these
two forces, the so-called pairing plus quadrupole interaction, has been perhaps
the most widely used simulation of nuclear interactions in heavy nuclei in the
last couple of decades. It can be written as:

where 0is the angle between the radius vectors to each particle (see Fig. 4.8)
and K-is the strength of the quadrupole part. Clearly, while Kpair is short-range,
the quadrupole component simulates at least part of the long-range aspect of
the residual interaction (see the following section).
The popularity of the pairing interaction, or of any other that reproduces
the low energy of/ = 0 coupled pairs of nucleons, clearly lies in the fact that all
even-even nuclei have 0+ ground states. The quadrupole interaction is moti-
vated, empirically, by the fact that nearly all nuclei more than a few mass
numbers away from closed shells display properties that can be described in
terms of quadrupole distortions of the spherical shape. Figure 4.13 shows the
differences between the pairing and 5-function interactions. For the case of
identical nucleons in equivalent orbits, both produce a low-lying 0+ state and a
grouped cluster of states with higher angular momentum. With the pairing
interaction, this latter group is unaffected, remaining degenerate. The 8-
function interaction seems to be a better approximation, as is clear in Fig. 4.5,
as well as Fig. 2.6 for the Sn isotopes. For many-particle configurations, the
quadrupole force dominates both the pairing or «5-function interactions and
the differences in their properties are washed out.

4.4 Multipolc Decomposition of Residual Interactions

It is useful at this point to consider a more general approach to calculating
nuclear interactions and their effects on energy levels in two-particle configu-
rations. This will help us understand, in a simple manner, the different effects
of different residual interactions. It will lead to a deeper appreciation of the <5-
function results, in particular the origin of the 2 curves in Figs. 4.9-4.12, and it
will shed more light on the relation between the pairing and 5-interactions. We
can expand any interaction that depends on the separation of the two particles
as

where the are given by

In the following we do not consider the effects of antisymmetrization explic-

 The Shell Model: Two-particle Configurations 95

itly, but the results are still valid for identical particles and for p-n systems
involving different major shells. If we write the wave function as a product of
radial and angular parts, we can then achieve a similar separation of the energy
shift into radial and angular parts and, moreover, separate the angular parts of
particles 1 and 2. We obtain, as in Eq. 4.5

where

and

As before, FR is a purely radial integral and depends on the details of the

specific interaction chosen, while Ak depends only on the angular coordinates
and is therefore completely independent of the interactions involved.
We now note a significant point. The summation in Eq. 4.24 is limited to
even k values from 0 to min (2jv 2/2,2lv 2/2). These limitations arise simply by
angular momentum conservation in the reduced matrix elements of Yk and as
specified by the triangle conditions involving the 9 -/ symbol. This severely
constrains the number of multipoles that go into the calculation of any given
interaction. For example, for any interaction V(il - r2) in a (d3/2 f 7/2 )
configuration, kmsi = 2 and hence k takes on only the values 0 and 2. Regardless
of the interaction, then, the energies of the four final states with / = 2, 3, 4, 5
are given in terms of only two equations. Indeed, as discussed by de Shalit and
Feshbach, one can show that

independent of the interaction. This case of d^, L,a particles should actually
describe the lowest states of 1738CL,r The experimental spacings are in the
ratios 1:1.1:2.0. They have the same sequencing of levels as in the calculation
(2-, 5", 3~, 4~), but the magnitudes of the spacings disagree. Without ever having
specified an interaction, or evaluated any radial wave functions, we can
unambiguously conclude either that an accurate description of these 38C1 states
requires more complex configurations, or that the interaction depends on
something other than (r}- r2) (e.g., spin). Note that this is the same nucleus
that we treated earlier with a 5-function interaction (Fig. 4.7). The earlier
results are different than these precisely because we used different strengths
(i.e., a spin dependence) for the S = 0 and S = 1 terms.
Equation 4.24 applies to any interaction that can be written in terms of the
separation of the two particles, that is, as V( \ ra - r 2 1). However, one some-
times encounters spin-dependent interactions. Then we obtain a result similar
to Eq. 4.24, with the same limitation on kmax, but including odd k. This point is
very important, since the 5-interaction can be shown to be equivalent to an
interaction V( \ ^ - r 2 1) times a spin-dependent operator. Thus, a ^-function
interaction is equivalent to an odd tensor interaction, even though its mul-
96 Shell Model and Residual Interactions

tipole expansion contains only even k. The reason is beyond the scope of our
treatment but involves the fact that the interaction takes place only at "con-
tact," and therefore a 5-function interaction can be multiplied by various
"exchange" interactions that have odd tensor character. This class of interac-
tions is crucial in nuclear structure since they have the special property of
conserving seniority (see Chapter 5).
For a 5-interaction, v^r ) is given by

Note that, despite its short-range, the 5-interaction has a monopole compo-
nent (k = 0) corresponding to a part of the interaction that is independent of
angle [P0(cos0) = constant]. This is a long-range piece par excellence and it
says that a "contact" interaction has a component that pays no heed to the
angular separation of the particles! Of course, since the monopole part is
constant over all (angular) space, it is the same for all relative orientations of
j\ and j2, that is, all / values. Hence, it does not contribute to splittings of a
multiple! but does give an overall shift to the multiplet. In practical calcula-
tions, the situation in regard to allowed multipoles can be even more bizarre.
Consider a 5-interaction between two d^ particles. Not only does the k = 0
multipole contribute, but the triangle conditions on k limit the multipoles to
only the values k < 3. No really short-range multipoles appear.
This seemingly paradoxical situation of an infinitly short-range interaction
being simulated only by relatively long-range multipoles is actually easy to
understand. This understanding reveals much about the relationship between
forces and the orbits and wave functions of the interacting particles. More
accurately, it clarifies the way particles from different orbits can "probe"
different interactions. It relates in a general philosophical way to how one
determines structure in any physical system. The general rule is that in order
to sample the structure of a given scale, the probe must be comparable to or
smaller than that scale. (It is difficult to distinguish between a potato and a
carrot by bombarding either with a truck.)
Imagine a proton and a neutron in an s1/2 orbit. They have only one m state
and are spherically symmetric wave functions: the orbit is uniformly spread
out, at a given radius, over time, in a spherical shell. The two particles are
therefore always in "contact" and are simply unable to sense any details of the
residual attraction. To them, a 5-force is identical to a constant force over all
space. The higher the; value of the interacting particles, the more sensitive
they are to the details of the force simply because higher; values have more
magnetic substates. This set of magnetic substates spans the same angular
range around the body of the nucleus and hence, each magnetic substate is
restricted to a narrower angular range. Thus, two particles in; - 13/2 orbits
coupled to / = 12 can sense the fine details of an interaction: each substate M
samples a different angular range of the force. In contrast, even though the
residual interaction may be a 5-force, two low; (e.g., s1/2, p1/2, PM,...) orbits
cannot "know" this. They are the wrong probe.
Fig. 4.14. Multipole decomposition of residual interactions: (a) Legendre polynomials Pk(cos &) plotted against 8 and J value for a (9/2,11/2) configuration, (b)
Contribution of different multipoles to AE (gn \mf) with a § function residual interaction. The short-dashed, and solid curves correspond to the multipoles k = 0,2
and k = 0,2,4,6, and 8, respectively. For the complete (k = 0-8) case, the even and odd J states are separately connected by broken lines, (c) Effect of the addition
of each successive multipole on the energies of an (hj^)2 configuration. (Parts (b) and (c) are based on Heyde, 1989.) (d) Effect of forces of different ranges on
the relative energy shifts as a function of 6, that is, of /, for two different two-particle configurations. (Schiffer and True, 1976.)
98 Shell Model and Residual Interactions

Coupled with the practical historic fact that most shell model calculations
have dealt with light nuclei and, therefore, low / states, we now see why
different interactions are often used to account for the same data, and why it
often is difficult to determine the details of the interaction.
The other extreme of k values (multipole order) can also be discussed.
Neglecting the limitation given by the jjr J212 quantum numbers (or, equiv-
alently, assuming a large j shell), we can relate the largest relevant k values to
the range of the force. Consider the classical picture shown in Fig. 4.8, where
0is the angular separation of the two particles. The distance between the two
particles is given by r2 = rf + r* + 2 r/2 cos 6. If we approximate the range by
a single number rr, the integration over 0 will be limited to 9 < rjr. The
Legendre polynomial Pk (cos#) oscillates more rapidly as k increases (see Fig.
4.14a). If Pk oscillates many times within the allowed integration range, the
integral will be small because of cancellation effects. In practice, therefore, for
a given force range, k is limited to value satisfying rik « rlr. Clearly, then, as
the range decreases, more and more A: values are required. The limiting case is
of course, the 5-interaction, although other factors (/,/ values) relating to the
ability of a given wave function to "sample" an interaction will come into play
to limit the allowed range of k values.
The idea of the multipole expansion of a residual interaction can be used to
obtain some earlier results in a physically transparent way that offers new
insights. This idea can also be used to derive a famous result known as the
parabolic rule for energies of states in p-n multiplets. We have already seen
how one can understand the /-dependence of the residual interaction in terms
of the angle between the two orbits involved (see Eqs. 4.18, 4.19 and Figs.
4.9-4.12). We can apply similar arguments to specific multipoles.
Each multipole, k, has an angular dependence Pk(cos6). The lowest few are
P0 = constant, Pz = l/2(3cos20-1), PA = l/8(35cos"0- 30cos20 + 3). The Pk were
shown in Fig. 4.14a. As Eq. 4.17 indicates, for a given j\,j2 there is a specific
relation between the total angular momentum 7 and the angle 6: thus, one can
plot Pt(cos0) equally well against J or /(/+!) (as is sometimes done). This
alternate scale is included in Fig. 4.17a for the example of/^ 9/2, ;2= 11/2 (e.g.,
a
(SwAin) configuration).
Obviously, since each multipole is proportional to P^cosfl), the interaction
in the kth multipole is strongest when Pk is largest. Thus, we can tell directly
which spin states will be most affected for each multipole.
The monopole (k = 0) component is constant, affects all states equally, and
simply gives an overall shift to the entire multiple!. This is an important effect,
altering the relative excitation energies of different multiplets, but it contrib-
utes nothing to the splitting.
It is the principal origin of the mass dependence of single-particle energies
that we discussed at the end of Chapter 3. In particular, the monopole p-n
interaction leads to the changes noted there (Fig. 3.4) in neutron single-
particle energies as a function of proton number and vice versa. In this
chapter, however, we are focusing on energy splittings due to residual interac-
tions and so we turn now to other multipoles.
 The Shell Model: Two-particle Configurations 99

The quadrupole (k = 2) component is strongest when 0=0° and 180° and

actually changes sign around 90°. (Recall that there is an overall minus sign to
be applied to Fig. 4.14a for an attractive interaction so that negative values on
the figure refer to repulsive effects.) Thus, a quadrupole interaction lowers the
energies of the extreme / values the most, but can in fact raise the energies of
intermediate spins. The general behavior of the quadrupole term is similar to
the ]\+J2+J even curve in Fig. 4.11 but this is somewhat accidental and we shall
return to the relationship between these results in a moment. We can easily
derive the specific dependence otAE(jlj2J) on/ for a k = 2 multipole. The AE
are just proportional to P2(cos6>) and, by Eq. 4.17,

where A, B, and C are functions of jv jv but are independent of /. This is a

parabola in /(/ + !) and is the parabolic rule discussed frequently by Paar. By
differentiating we can find the vertex, J\

and, solving the quadratic,

Note that Jy is normally the highest-lying member of a downward pointing

parabola since the overlap of the particles is largest for Jnax and J^. In the
g9/2hn/2 case in Fig. 4.14a, this gives Jv« 7, which is indeed where the P2(cos6>)
minimum occurs. For this multiplet, the / = 1,10 states are lowest,/~7 highest.
The relationship of these ideas to the /-dependence in Figs. 4.9-4.12 is
informative, particularly the split into two curves*. In general, points falling
alternately (for odd and even /) on these two curves cannot be described by a
single parabola in/(/+1) (we will see one exception later). To generate the 8-
function results, the higher multipoles must be added in. From Fig. 4.14a it is
clear that these oscillate more rapidly in 9 or J, and this introduces an
irregularity or zig-zag pattern to the spin dependence. For example, a hexade-
capole component would raise E(7) and lower that of the J= 4 and 9 states. It
also broadens the minimum and tends to shift it toward the spin 8 level. A
P6(cos0) term will add further perturbations. When all the multipoles are
summed, with amplitudes appropriate to the 5-interaction, two curves similar
to those of Fig. 4.11 are reproduced.
This is evident in Fig. 4.14b, which shows the strength of the interaction as
a function of J for different combinations of multipoles (with amplitudes
appropriate to a 5-function). For the I g9rahu/27) configuration, the k = 0 + 2
curve is smooth and minimizes at J = 7, as we have just seen. If this curve is
replotted against J(J + 1), an exact parabola is obtained. Adding in all the

*I am indebted to K. Heyde for much helpful advice in this section and for Figs. 4.lib and 4.14c.
100 Shell Model and Residual Interactions

multipoles, k = 0,2,4,6,8, gives the solid curve. If we connect just the odd J
states along this curve, we get a dependence just like the jl+J2+J odd curve of
Fig. 4.11. If we connect the even/values we get the j\+j2+J even curve of Fig.
4.11. Thus we see that the separation of the two curves comes from the higher
multipoles and reflects the Pauli effects since these multipoles have shorter
ranges.
Figure 4.14c shows this in a particularly illuminating way for the identical
particle I (h]1/2)27) configuration. Here, the energies obtained with the addi-
tion of each successive multipole in a ^-interaction are displayed in level
scheme form. For k - 0,2 only, a parabolic behavior is observed: the 0+, 2+, 4+
levels are lowest, the intermediate spins 6+, 8+ highest and the largest spin, 10+,
lower again. As shorter and shorter range multipoles are added, the levels
shift toward the characteristic 5-function sequence shown on the right. We
have discussed how such a sequence, particularly the weak effects of a 5-
interaction for the Jmsa states (since they have S = 1), is a specific Pauli principle
effect. Now we see exactly how this comes in through the shortest-range
multipoles, which are in fact repulsive for these spin states, reflecting the Pauli
prohibition against "contact."
Figures 4.14a and 4.14c also dramatically show why the 0+ state (or, gener-
ally, the lowest state of a given multiple!) is lowered so much. This state
corresponds to orbit planes closest to 9 = 180° (or 0°, depending on the; values
if nonequivalent orbits are involved). Here, all the multipoles contribute
coherently because P^cos 180° or 0°) is always unity. Only for other spins do
cancellation effects enter.
Finally, we can use Fig. 4.14c to better understand the relation between 5-
and pairing forces. Imagine continuing this figure further to the right. The
0+ - J separation would grow and the separation among the J* * Q+ states
would diminish relative to their separation from the 0+ level. The limit is the
pairing picture shown in Fig. 4.13. Although we do not write the pairing
interaction in terms of multipoles, we see that, in effect, it corresponds to the
dominance of very high ones. This, incidentally, is the origin and basis for the
phrase occasionally encountered that the pairing force is of even "shorter
range" than the 5-interaction, a statement that sounds paradoxical.
We have now discussed, in several ways, the relation between multipoles
and the (angular) ranges of the forces they describe and their effects on differ-
ent J states of two-particle multiplets: briefly, attractive short-range (higher
multipole order) forces tend to lower especially the Jma and Jain states (subject
always to Pauli princliple constraints). We can see this relation even more
explicitly by using forces that are finite (radial) range. To this end, Fig. 4.14d
shows the energy shifts calculated for two configurations for a Yukawa type
force with range parameter r = 0.1,1.0 and 5.0 fm. In general, the shorter the
range, the stronger the effects on the extreme J (or 0) states where the orbits
are most nearly coplanar and the particles, on average, closer. The different
patterns for r = 0.1fm for the 7=1 interaction just reflect the cot 9/2 or
tan 8/2 dependences discussed for the two different configurations (see Table
4.2). This difference, a Pauli effect, is washed out for longer-range forces, as is
 The Shell Model: Two-particle Configurations 101

most of the J (or ff) dependence, since the force becomes rather insensitive to
the separation of the particles within the nuclear volume.
With this discussion of multipole contributions and force ranges, the reader
should now be in a position to estimate, at least qualitatively, the effect of any
particular interaction in a given configuation with little or no explicit calcula-
tion.
An interesting special case of the multipole decomposition is that of a
(PM^M) P~n configuration. For this configuration, Eq. 4.24 only allows k = 0,
2 values, and the k = 0 multipole is irrelevant for splittings, so the 5-interaction
in this case should give exactly the parabolic dependence on /(/ +1) character-
istic of P2(cos9). Yet, from our previous discussion, we know that the J = 0,2
(/i + J2 + J odd) states would occur on the upper curve in Fig. 4.11 and the
J = 1, 3 states on the lower curve. Although this hardly seems to yield a
parabola, careful inspection of the curves in Fig. 4.11 for the angles appropri-
ate to the | dM pM /) configuration shows that these four points have an exact
parabolic form against /(/ +1).
Actually, this result suggests why a quadrupole interaction is more impor-
tant than it might at first seem: because of the angular momentum constraints
on k , many higher multipoles that might normally contribute are elimi-
nated. Moreover, the 5-function is not necessarily the best choice of interac-
tion. Other, "finite range" interactions (e.g., Gaussian e~°(l r l ~ r z D ) are often
used and have relatively larger low-fc amplitudes. Finally, there are often other
residual interactions besides the p-n interaction, such as particle-core cou-
pling contributions that are dominated by quadrupole components. Indeed,
we saw in Section 4.2 that analyses of empirical two-particle multiplets do
suggest evidence for enhanced quadrupole interactions. (We shall see in
Chapter 6 that quadrupole core "vibrations" are the dominant low lying
"collective" modes in nuclei.)
Thus, the parabolic rule is often an excellent approximation. Figure 4.15
shows a few examples taken from Paar and introduces one final but important
point. The (g7;2d3/2) multiple! of J = 2*- 5+ states in 1Z2Sb is well reproduced by
a simple parabola in /(/ +1), as is the (g7/2h11/2) multiple!. 48Sc and 116I also show
multiplets with beautiful empirical parabolic behavior, except that they are
inverted! The reason is well understood. In 2148Sc27, the (iri2^1/2n') multiple!
really a particle-/io/e p-n configuration (ilrip f \/2n) = (iirip f~l7/2n), as is Ihe
(gra/'hj^) configuration in 49116In67. In this case, the residual interaction has
the opposite sign of a particle-particle or hole-hole multiple!. In other words,
it is repulsive, and the J states with high p-n overlap (/„,„, J^J are raised in
energy, while Jn is lowered the most. We note that this change in sign is not
always the case; in Chapter 5 we will see i! as a charac!eristic of even multipole
interactions such as the quadrupole interaction we are considering here. Inler-
estingly, it does not apply lo the ^-interaction (even though we have discussed
its expansion in even multipoles). For odd multipole interactions AE(p-p) or
AE(h-h) is identical to AE(p-h).
In Chapter 5, we will discuss at length the pairing interaction and the
concept of the quasi-particle, which is a state only partially occupied, neither
102 Shell Model and Residual Interactions

Fig. 4.15. Illustrations of the parabolic rule for a quadrupole residual interaction for several p-n
multiplets(Paar,1979).

fully particle nor fully hole. Anticipating that discussion, the occupancy of an
orbit is given by a probability, denoted V2. (the number of particles in the orbit
is (2; + l)^ 2 ). The orbit emptiness is U1. and U2.+ V2 .= 1. Thus, we can rewrite
Eq. 4.28 for the general situation of quasi-particles as

Thus, for a given proton number (particle "1"), (t/2, - V 2 ) is constant but
(f/2z - V22) changes from +1 to -1 as the neutron orbit ("2") is filled over a
 The Shell Model: Two-particle Configurations 103

sequence of isotopes. Therefore, the parabolic rule actually refers to a whole

family of possible parabolas for given jv j2, ranging from bowl-shaped (p-h
case) to nearly flat to a convex upward parabola (p-p or h-h cases). The
theoretical shapes, for several values of VMmn in a (gr1,^ h^) multiplet are
illustrated in Fig. 4.16.
It is worth mentioning two applications of these ideas. The fact that
different interactions have different multipole composition can sometimes be
used to gain information on the nature of the effective interaction. This has
been extensively pursued by Schiffer and Molinari and co-workers, who, for
example, deduced effective multipole coefficients applicable to broad ranges
of nuclei, and by Heyde and co-workers who studied Gaussian and other
interactions. In general, stronger quadrupole components are required than
given by a 5-force, signaling the need for "finite" range interactions.
Secondly, this geometric interpretation of the interaction can provide
important intuitive clues to the configurations in certain states without the

Fig. 4.16. Dependence of the shape of (g'lm hu/2n) multiplet parabolic splittings as a function of
the occupancy. The uppermost curve corresponds to a nearly full hn/2|i orbit and therefore a
hole-hole configuration. The lowest (h-p) curve corresponds to an hn/2ji orbit with only a few
particles and therefore a hole-particle configuration (based on Van Maldeghem, 1985).
104 Shell Model and Residual Interactions

need for prior detailed calculations. For example, let us compare the multiplets
(hn/2)2 and (dM)2. Clearly, from Eq. 4.17, the angle between the orbits in the 2+
state is less in (h,M)2 [6 = 24° compared to 49° for (d^2]. Therefore, the
overlap is greater and the 2+ energy lower so that the 0+- 2+ spacing will be less
(see Fig. 4.3). By comparing 0+- 2+-4+-... intervals, one can sometimes deduce
evidence for particular / components of the wave functions. Care must be
taken in this, however, since 2+ states, even in singly magic nuclei, often have
several significant components, exhibiting "collective" behavior and lower
energies than can be predicted for single two-particle configurations.
Nevertheless, changes in spacings of yrast states (the first states of each J
value) across a sequence of nuclei can sometimes signal the emergence of
particular shell model components.

4.5 Some Other Results (Average Shifts, Hole, and Particle-Hole

Configurations)
In closing the discussion of two-particle systems, we note a few other impor-
tant results. Consider the case of two identical particles in equivalent orbits
and an odd tensor interaction. Only odd k terms appear in Eq. 4.24 in this case.
The total summed interaction energy A£tot is given in general by

From the properties of sums over 6 - / symbols, without ever considering the
radial functions /*, we find that by restricting ourselves to odd tensor interac-
tions,

hThat is, the total shift of all the levels in a configuration of two identical
particles in the same orbit is simply related just to the shift of the J- 0 level, for
any odd tensor interaction. Since the 5-function interaction is equivalent to an
odd tensor interaction, it can be expected that Eq. 4.31 will be approximately
satisfied in real nuclei. Finally, we re-emphasize that it is the separation into
angular and radial coordinates that allows results that are totally independent
of the details of the interaction (except, in this case, for the assumption of odd
tensors).
Thus far, we have restricted the discussion to two valence particles. Multi-
particle configurations will be dealt with shortly. It is useful, however, to
consider one simple result that allows a nice generalization of the present
discussion. Using the formalism of the seniority scheme, it is possible to relate
the matrix elements of any two-body operator in the/" configuration to matrix
elements in configurations with fewer particles. For the large class of interac-
tions that conserve seniority (including a 5-function that is a prototype of
short-range interactions) it can be shown that in a multiparticle, f configura-
tion, the more nucleons that are paired off to J - 0, the lower-lying the state will
be. Thus, for n even, the lowest state of the;'" configuration is a J* = 0+ state
with all nucleons paired to 7 = 0. (The &E(j^) for the levels of the configuration
 The Shell Model: Two-particle Configurations 105

\f-J) is just a special case of this). This result is sometimes called the pairing
property, and is characteristic of any interaction that conserves seniority.
Another key result is that the order and spacing of the / * 0 states in the /"
configuration with two unpaired particles will be exactly the same as in the j2
configuration. The more realistic situation in which the valence nucleons
occupy several; orbits, jv ;2>... in a multiparticle configuration I/T1,/'"2,- • • /)
with n; even, is just a generalization. By the first result, the / = 0 combination
will lie lower than one in which a pair of nucleons is coupled to / * 0. Thus we
can generalize these conclusions and state that the ground state of even-even
nuclei will always be 0+. Moreover, since the odd J levels of I ;"7) are unaffected
by a 5-interaction (n even), the low-lying positive parity states should all have
even spin and should increase in energy with spin. Remarkably, in view of the
simplicity of the argument, this is an almost universally observed situation.
Furthermore, we noted that in n=- two-particle configurations of identical
particles, I/^T), it is the odd spin levels that are lowered. A simple generaliza-
tion using the seniority scheme shows that the lowest-lying negative parity
levels in even-even nuclei should have odd spin. This is also almost always
observed, and is known as the Talmi-Glaubman rule.
One can go even further. Consider an arbitrary shell in heavy nuclei
consisting of several states (/ orbits) from one shell and one unique parity orbit
from the next shell. Recall from Fig. 3.2 that the highest normal parity /value,
/roax, in each succeeding major shell increases by one. The corresponding
highest j value is jma[ = /max -1/2. (The ;' = /maj +1/2 is the orbit brought down by
the spin orbit interaction into the next lower shell.) The next highest / is
lma - 3/2. The unique parity orbit is the / +1/2 coupling from the next highest
shell, and so has ;unique = /mai + 1 + 1/2 = /max + 3/2. For example, in the 50-82
shell, /mM = 4, so the highest normal parity orbits are the g7/2, ds/2 and the unique
parity orbit is h11/2. In the g^ - hn/2 configuration, our rules (see Table 4.2)
show that the /=;',+ /2 = 7/2 + 11/2 = 9" state is brought lowest and the 2~ is high-
lying at the unperturbed energy. In the d5/2-hn/2 case, the J" = 3~ is lowest. A
little thought shows that this relationship of/ values is generally true for any
shell in heavy nuclei. Therefore, we can obtain, without calculation, two
general predictions. First, the lowest-spin, negative parity, two-particle state
in an even-even nucleus will be a 2~ level: there is no simple spin combination
that gives a 1" level. However, since this 2~ level always results from a/ = M72
normal parity orbit coupled antiparallel to a ;' = I + 1/2 unique parity orbit, it
will always occur rather high in energy. Therefore, the lowest-lying, low-spin,
negative parity state will be the 3" level. This prediction is borne out in nearly
all heavy even-even nondoubly magic nuclei (in closed shell nuclei two differ-
ent shells are involved in n= - excitations). Examples were shown in Figs. 2.5
(210Pb) and 2.6 (Sn). The main exceptions to all these predictions are them-
selves illuminating: they concern nonspherical, or deformed, nuclei in which,
to borrow terminology from a later chapter, one sometimes encounters two-
quasi-particle, negative parity excitations involving a unique parity Nilsson
orbit (e.g., the low-lying 4- state in 168Er). But, as is well known, the nonspheri-
cal character of these nuclei is itself induced by a strong residual quadrupole
interaction: being an even rank tensor, it does not conserve seniority and
106 Shell Model and Residual Interactions

therefore, a priori, one would not expect the preceding arguments to apply.
A final topic to deal with before considering more complex multiparticle
systems is the spectra of hole and particle-hole configurations, where, by the
former, we mean those of the type (/'-") = (/2'+1 -"). We have briefly mentioned
these in our discussion of multipole forces. Here, we obtain a few more explicit
results.
For hole states, a diagonal matrix element of any single particle operator
acting on an n-hole state \j~"JM) will be equal in magnitude to its value in the
n-particle system \j"JM). The sign relation will depend on the odd or even
tensor character of the operator. This is easy to see by considering all the
possible m states of the n-hole configuration. There will be n unoccupied m
states. The expectation value of a single-particle operator Om (which we
denote O for simplicity) will be the sum of its expectation values over all the
particles:

where or" simply indicates the n-hole state, but where the calculation is carried
out over the 2; + 1 - n particles. This is clearly equal to

that is, equal to the sum over the entire shell minus that for the n missing
particles. But, we have shown above that the first term must vanish since the
closed; shell can have no preferred direction in space and, therefore

Note that the n-particle and n-hole states are not exactly equivalent. They
have different total M values. For n particles in a given / state occupying m
states ml,mv...m ,M ="ZMm.. For wholes in the same orbit, one clearly must
have Mh = -Mf (e.g., 1/2 + 3/2 = -[5/2 + (-1/2) + (-3/2) + (-5/2)]). So the above
matrix element can be written (in simplified notation)

where the subscripts indicate the n-hole and n-particle configurations. Using
the Wigner Eckart theorem for a tensor operator O*of rank k gives

and hence we can now relate the matrix elements of any single-particle opera-
tor in states of the same JM for n-hole and n-particle configurations:
 The Shell Model: Two-particle Configurations 107

This gives the critical result that the expectation value of any odd tensor single-
particle operator (e.g., magnetic moment) in an n-hole state is the same as the
corresponding n-particle expectation value, while for even rank tensors (e.g.,
quadrupole moment), the two expectation values are the negatives of each
other.
In Chapter 5 we shall encounter more general results of this special case.
This has immediate consequences of great importance. It implies, for ex-
ample, that magnetic moments or dipole transition rates are the same for
corresponding particle and hole configurations, while quadrupole moments or
E2 matrix elements change sign. Therefore, such matrix elements must vanish
at mid-shell. Of course, these features are well known empirically, although
they are partly obscured by configuration mixing effects.
A nice extension of this is to the interaction energies in particle-hole
configurations relative to those in particle-particle ones. Compare a
proton-neutron particle-particle configuration with a proton hole-neutron
particle configuration. The interaction is a product of proton and neutron
tensor operators. The effect of the neutron operator on the neutron wave
functions (a particle in both cases) is the same for both configurations. Only
the proton operator acts on different configurations (particle in one case, hole
in the other) and the above results for one-body operators then apply. Thus,
for an odd tensor interaction (product of odd tensor one-body operators), the
energy shifts AE(j -';„ /) = A£(/p jnf) while, for an even tensor interaction,
AE(jf-ljn /) = -A£(yn /). The U, independence of a quadrupole force leading
to the up-and-down pointing parabolas for p-p (or h-h) and p-h configura-
tions in Eq. 4.30 is an example of this sign change. For the more general case
of a mixed-interaction, the results are slightly more complicated and are given
by

Note that this gives the energy differences for different J values. The energies
themselves can have a constant additive term that is/-independent (but that
can depend on/). Thus, while Eq. 4.34, or the simpler relations for odd and
even tensor interactions, relates p-h and p-p spectra for a given j shell, it does
not relate such spectra for different; orbits in a major shell.
Since, physically, any very short-range interaction cannot give results quali-
tatively different from that of a 5-function (equivalent to an odd tensor
interaction), one expects that the odd tensor result will be closer to that
observed empirically. In other words, the energies (order and spacing) will be
identical in the p-h and p-p configurations. Nevertheless, Eq. 4.34 is very
important because it is so general. It is valid for any two-body interaction in a
//-coupling scheme. Moreover, it has a practical use from a different, empiri-
cal, point of view. Without knowing anything about the actual interaction, the
knowledge of the order and spacings in a p-p configuration allows one to
predict those in the corresponding p-h configuration. A classic example of this
is the comparison of ,738C121 and 1940K21: the former is expected to be a (dMpf7;2J
configuration, while the latter is (d^f^J = (d'^f^J. The comparison is
108 Shell Model and Residual Interactions

included in Fig. 4.7, where the empirical "°K energies were used to calculate the
level energies in 38C1. The agreement is excellent.
Finally, we consider the interaction energies of the (j*f) configuration
relative to those in the (j"J) configuration. By arguments similar to the
preceding in the m-scheme, it is easy to show that the interaction energy for any
two-body operator in an n-hole state will be equal to that in the «-particle state
plus an additive constant. Therefore all spadngs in n-hole configurations
\i~nJ) should be identical to those in the corresponding n-particle states \ff).
Note an important point analogous to a comment we stressed in regard to the
effect of closed shells on open shells in Chapter 3. As used here, the term shell
refers to a single ; shell, not a major shell consisting, generally, of several;
values. The additive constant relating «-hole to n-particle configuration need
not be independent of/. Therefore, the full set of spacings in a configuration
such as jini jl "2 ---j] will not generally equal those in j\l j"2 • • • J\. The equal-
ity of spacings applies to each /, individually. If the realistic wave functions
contain admixtures of several / values, there is no simple general relation of
particle and hole energies.
A final point: thus far we have only discussed diagonal matrix elements
(energy shifts) due to a 5-interaction, but off diagonal effects are equally
important. These, of course, give rise to the mixed wave functions character-
istic of realistic shell model calculations. Expressions for the off-diagonal
matrix elements (JJ2\S\ /3;4) are similar to Eq. 4.7, except there is now a
second 3 -;' symbol involving/,, ;4. The strength of a given matrix element
then depends on the angular correlation of the two particles in both initial and
final states, as well as on the radial overlaps.

It is worthwhile giving a perspective on what we have done in this chapter.

We first considered a 5-function residual interaction in a rather quantitative
way, for both like and unlike particles. Then, we noted the geometrical
relation of the orbits in two-body interactions. With a corresponding geomet-
rical understanding of the force between these two particles, it is then trivial to
deduce (intuitively or quantitatively) the effects of that force on the various
total angular momenta J. We did this in detail for the 8-interaction, recovering
the earlier results. We then developed the idea of a multipole expansion of any
central interaction that gives the needed geometrical structure, and showed
how it can be used to estimate the effects of an arbitrary two-body residual
interaction of such type. All this was carried out for two-particle configura-
tions. Now we turn to see how these results can be used for the multiparticle
case.
 5
MULTIPARTICLE CONFIGURATIONS

We now turn to a more systematic treatment of multiparticle (n > 2) configura-

tions of valence nucleons. We have already anticipated a few results from this
discussion in Chapter 4. Without having done so, the two-particle results
would have been completely useless. We needed to understand, for example,
the result that closed shells could be ignored in considering residual interac-
tions, along with some of the results relating p-h to p-p spectra. This allowed
us to apply the two-particle discussion to the case of two valence particles or
holes relative to any closed shell. We also quoted the enormously important
conclusion that (if seniority is conserved) many predictions for | ff) configura-
tions are identical to those for \fj) configurations. (The requirement of
seniority conservation basically means that the other (rc-2)-particles couple
pairwise to Jn 2 = 0.) Thus the two-particle results could be used for nearly all
spherical nuclei.

5.1 /Values in Multiparticle Configurations: The m Scheme

Clearly, our first task in the systematic study of multiparticle systems is to
consider which values of the total angular momenta / are permissible for an n-
particle system. The principal consideration here is, of course, the restrictions
imposed by the Pauli principle. There are several ways of approaching this
issue. The most physically transparent and easy to use is the so-called ra-
scheme, to which we now turn. Let us start by working out an explicit case.
Consider three identical nucleons in a ds;2 orbit. They cannot couple to a
J = 15/2, since such a state must have an M -15/2 substate. The only way such
a substate can be made is by placing all three particles in m = 5/2 substates,
violating the Pauli principle. Similarly, J = 13/2 or 11/2 states are impossible
since the former would require two of the particles to be in m = 5/2 substates
and / = 11/2 would require two particles in m = 3/2 states. However, a J = 9/2
state is indeed possible since it can be formed by placing three particles in the
states//n, = 5/2,5/2,;2m2 = 5/2,3/2, }3m3 = 5/2,1/2.
It is easy to see that a slight generalization of this example gives the
following result: in the configuration;'" the maximum angular momentum is
given bv ,

109
110 Shell Model and Residual Interactions

Table 5.1. m scheme for the configuration I (7/2)2J)*

A=7/2 ;2=7/2
m, m
2
Af /
7/2 5/2 6"
7/2 3/2 5
7/2 1/2 4
7/2 -1/2 3 6
7/2 -3/2 2
7/2 -5/2 1
7/2 -7/2 0.
5/2 3/2 4"
5/2 1/2 3
5/2 -1/2 2 4
5/2 -3/2 1
5/2 -5/2 0
3/2 1/2 2"
3/2 -1/2 1 2
3/2 -3/2 0.
1/2 -1/2 0] 0
"Only positive total M values are shown, the table is symmetric for M < 0.

For the (ds/2)3 case just discussed, this gives Jnm = 3; - 3 = 9/2.
Thus, we see an example of how a consideration of the possible m substate
occupations can give information on permissible total angular momenta J.
Basically, the m-scheme is a systematic set of procedures for doing this in a
general case.
The m-scheme is best described with a detailed example. Earlier we
showed that only even J values were allowed in the j1 configuration of identical
nucleons. This was argued in terms of the symmetry properties of spherical
harmonics. However, it can also be seen by inspecting the possible magnetic
substates. Table 5.1 summarizes the allowed magnetic substates for a two-
particle configuration (7/2)2. (Actually, the M < 0 cases are omitted since they
are completely symmetric to the Af > 0 cases.) We construct such a table
starting with the highest magnetic substate for particle 1 and list all of the
possible substates for particle 2 allowed by the Pauli principle, then carrying
out the same procedure for the next lower magnetic substate (in this case 5/2)
for the first particle. It is necessary to recall that the two nucleons are indistin-
guishable so a combination 5/2,7/2, for example, is not allowed, since the 7/2
and 5/2 combination has already been listed. Continuing in this way, we obtain
all of the possible m values for the two-particle system. Since a given total
angular momentum / must have magnetic substates M = J,J- !,...-(/ -1), -J,
it is now trivial to deduce the possible final / values using the table. In the
present example, there must be a / = 6 state, since there is a M = 6 configura-
tion. This 7 = 6 state has magnetic substates M = 6, 5, 4, 3, 2, 1, 0, and the
corresponding negative values. Thus the top seven magnetic substates listed in
the table must be used up for this single / = 6 state. There is no M = 5 magnetic
substate left over and thus there cannot be a J = 5 state. There can, however,
be a / = 4 configuration that consumes the M = 4,3,2,1,0 magnetic substates
 Multiparticle Configurations 111

Table 5.2. m scheme for the configuration I (5/2)V)

A = 5/2 A=5/2 A = 5/2 M /
m
i IWj m3

5/2 3/2 1/2 9/2"

5/2 3/2 -1/2 7/2 /X
5/2 3/2 -3/2 5/2
5/2 3/2 -5/2 3/2
5/2 1/2 -1/2 5/2
5/2 1/2 -3/2 3/2 < 5/2 9/2
5/2 1/2 -5/2 1/2.
5/2 -3/2 1/2"
/x
-1/2
3/2 1/2 -1/2 3/2=
\ 3/2
3/2 1/2 -3/2 1/2.
7 = 9/2, 5/2, 3/2

"The full set of allowable m. combinations that give M > 0 are obtained by the conditions nij > 0, m3 < m2 < mt
and no two m.values identical.

listed next in the table. By continuing this argument, one sees that there is no
J = 3 state, but there are a / = 2 and a J = 0 state, thus proving in a different way
the result obtained earlier that two identical particles in the same orbit can
couple only to even total angular momenta /.
The case for (5/2)3 configuration is shown in Table 5.2, which will not be
discussed in detail although the reader may go through the example and verify
the results just as we did for the (7/2)2 case. Clearly, for multinucleon configu-
rations where «is large, this procedure can be lengthy. Other techniques are
available. However, the m-scheme is important because it shows in a transpar-
ent way how the physical effects of the Pauli principle arise.
We noted above that the m-scheme gives a rule for the maximum permissible
/value in a/" configuration of identical particles very simply. It also gives, in an
equally simple way, the result that a/1 configuration can never have a state with
/ = Jm!at - 1. We will show this only for the case of two particles, but the
generalization is straightforward. (Although the following considerations are
general, reference to the specific example in Table 5.1 will clarify the argu-
ments.) The maximum Jinaf configuration is /mM = 2; -1. A state J = /max - 1
would have (if it existed) J = 2(j -1). One value of M = Jmia - 1 must be used
for the J - /mM state. Therefore, in order to have a state with J = Jmax -1 = 2(/ -1),
there must be a second permissible M = Jma - 1 = 2(/ - 1) state. This cannot
involve a particle in an m = j state, since that state is already consumed for the
/ = /max level. Therefore, the only way to make another magnetic substate
M - 2(j - 1) is to have two particles with m - j - 1. But this violates the Pauli
principle, and therefore is impossible, proving that a / = /mai - 1 state never
exists in a j2 configuration. As noted, this can be generalized to the ;'"
configuration.

5.2 Coefficients of Fractional Parentage (CFP)

Now that we have a feeling for those J values that can be obtained for any
multiparticle configuration, we can discuss the effects of various interactions
112 Shell Model and Residual Interactions

on the energies of such / states and dynamic matrix elements involving such
configurations.
Clearly, when considerations such as those discussed in Chapter 4 are
attempted for multiparticle configurations, the situation rapidly becomes
much more complex. To see this in a simple example, consider a (d5/2)3
configuration of identical nucleons. A J = 5/2 state can be made in three
distinct ways by first coupling two particles to an intermediate J' = 0,2, 4:

and

However, the (5/2)3 configuration has only one J = 5/2 state (see Table 5.2). Its
wave function must therefore be a totally antisymmetric linear combination of
these three basis states. The normalized coefficients in this linear combination
are called coefficients of fractional parentage (CFP): their squares give the
probability that a given final state is constructed from a specific "parent"
configuration—in this case, a two-particle state. The relative magnitudes of
the three CFP's for the (d5/2)3 configuration are not arbitrary, but are given by
certain angular momentum coupling coefficients. CFP coefficients can be
constructed not only for three-particle configurations, but for any n.
Unfortunately, the complexity of the possible couplings makes the notation
for the coefficients rather complex and this has deterred many nuclear physi-
cists from delving into the subject; the formalism can be terrifying. Formal
textbooks can be filled with page after page of long, daunting expressions in-
volving sequences of CFP coefficients, angular momentum coefficients, sum-
mations over them, and the like. Here, we attempt to cut through much of this
by summarizing some of the essential results and their motivations with a few
simple examples. This is not entirely satisfactory, since it deprives the reader
of an appreciation of the beauty and power of the formalism. Moreover, as the
author can personally attest, while a simple presentation of the final results
saves the reader from the tedium of struggling through their derivation, it also
confers on them an element of mystery—the reader is left with a sense of
wonder at how one can start with such general interactions and general
configurations and end up with very simple final results. He or she may glance
back over the imposing derivations in the hopes of seeing where some simpli-
fying assumption or some restrictive case has been invoked. The real power
and beauty of the method, however, is that such assumptions are usually not
required: very general results that enormously simplify the treatment of many-
particle shell model configurations can often be obtained.
In any case, we will introduce the formal notation for CFP coefficients, but
will avoid their manipulation as much as possible. We will derive one simple
result that illustrates their power and economy.
 Multiparticle Configurations 113

Consider a configuration of identical nucleons in equivalent orbits, the

jnaJM state (a specifies any additional quantum numbers needed to describe
the states). This n-particle state can be written in terms of (n-2)-particle wave
functions by using the two-particle CFP coefficients (there are also analo-
gously defined one-particle CFP coefficients). The CFP is denoted

and determines, as in the preceding three-particle case, the probability that the
wave function | fJM1) can be written in terms of the (n-2)-particle configuration
\j"^(Jn_2Mn 2)> coupled to a two-particle configuration |/172M2). The defining
equation is therefore

Clearly, then,

We can see how the concept of CFP coefficients and the parentage of n-
particle configurations is useful. Consider a configuration \ff) and ask what
the energy shifts, AE(jnJ), are for each final /value for an arbitrary interaction.
(Note that, as usual, we drop the magnetic quantum numbers to simplify the
notation.) First, we note that, since the particles are indistinguishable, the total
interaction energy in any final state J is given simply by the interaction energy
for any pair of nucleons (say, particles 1 and 2) times the total number of
possible pairs n(n - l)/2. However, the two-particle matrix element Vn can
only depend on 72, and not on the way in which 72 is coupled with Jn 2 for the
other n-2 nucleons to give the final /. The total interaction energy for particles
1 and 2 is just the sum of the interaction energies for each two-particle angular
momentum ]^ multiplied by the probability of each 72 in the state \ff). We
denote this probability W(fJJ2). Thus, we can immediately write the interac-
tion energy from particles 1 and 2 in the state I/1/) as:

where the W coefficient is the sum of the squares of the CFP coefficients for a
given J2 over all possible values Jn 2. That is,

The interesting point here is that there are in general fewer values of 72 than
there are of 7. For example, in the (7/2)3 configuration, J can be 15/2, 11/2,
9/2,7/2,5/2,3/2 while (7/2)2 can only couple to 72 = 0,2,4, and 6. Thus, by Eq.
5.2, the six energies of the configuration j"J are given in terms of the four
matrix elements
114 Shell Model and Residual Interactions

Fig. 5.1. Comparison of the low-lying empirical levels of 51V with calculations obtained by
coupling an (J/2 proton to an (f?ffl)2 two-particle configuration (right) and by coupling an f?/2 proton
to the empirical levels of 5aTi. (See deShalit, 1974.)

The beauty of this is that these matrix elements are usually easy to calculate
for a known interaction and, even when the interaction is not known, empirical
values for them can be obtained from the neighboring even-even nucleus
(with n = 2). This can then be used to calculate the energy levels of the adjacent
odd mass nucleus.
We have discussed the (7/2)3 example here because it is treated in detail in
de Shalit and Feshbach, where the low-lying (f7/2)3 energy levels of 51V are
calculated in terms of the empirically known (f7/2)2 levels of 50Ti (0+:0,2': 1.55,
4+:2.68, 6+: 3.2 MeV). The results are shown in Fig. 5.1; the agreement is
remarkably good for such a simple approach. Note once again that nowhere in
this discussion has any aspect of the interaction been specified, except to
assume that it is two-body only. We could also have calculated 51V with the
same formulas using a 5-function interaction to simulate 50Ti, that is, to define
the (f?/2)2 matrix elements. Normalising the 8-function strength to the 0'-6f
spacing in 50Ti gives calculated 50Ti energies of (M), 2+:2.68, 4+:3.0, and 6':3.2
MeV. These have a different distribution than the empirical levels and, when
applied to 51 V, give the fit on the right of Fig. 5.1. Clearly, this approach is not
 Multiparticle Configurations 115

nearly as successful. The point is that the empirical 50Ti spectrum automatically
includes all relevant interactions in the (f,^)2 system. The CFP techniques
relate this directly to51V, independent of a knowledge or guess of the interac-
tion. Thus, an understanding of the makeup of an n-particle configuration in
terms of its (n-2)-particle structure can greatly simplify the treatment of
nuclear spectra in complex systems. The present results can be generalized to
n > 3, and provide comparable, and even greater, simplifications.

5.3 Multiparticle Configurations j": The Seniority Scheme

When there are numerous particles outside closed shells, they can enter
different shell model orbits. For example, in 40"Zr59 the nine valence neutrons
might be in a configuration (dM)6 (g^)3. Here, the dM shell is filled and the ear-
lier arguments on the effect of closed shells on the values of AE( j2'1 +1 /"2J) tell
us that the dM orbit can be neglected, so this configuration is equivalent to
(g7/2)3. Now, consider 95Zr. In this case, the lowest expected configuration
would be (d5/2)5. By the particle-hole equivalency discussed earlier, this is
exactly equivalent to a single neutron in the d5/2 orbit, leading to a one-state
configuration with J =j = 5/2 and, indeed, the ground state of 95Zr is 5/2+.
However, one could also imagine excited states in 95Zr of the form (d5/2)3 (g7/2)2.
Normally, at least near closed shells, such configurations are rather high-lying
excited states: our primary interest is usually in the lowest-lying levels in which
as many particles as possible are packed into the lowest accessible j value.
Thus, at least in simple shell model treatments, one is frequently interested in
;" configurations. Moreover, even though realistic shell model calculations will
often involve important components coupling two / values, an understanding
of the single ;' case greatly helps to interpret and even anticipate such calcula-
tions. So far, we have ignored the possibility of both valence protons and neu-
trons. This clearly complicates the situation, as seen in the discussion earlier of
the 5-function interaction for p-n systems. Moreover, as we shall see later,
once one has nucleons of both types outside of closed shells, collective effects
rapidly accumulate and other models provide alternate, and often better,
approaches. Therefore, it is appropriate to again stress the/1 configuration of
identical nucleons. Despite this restriction, the following considerations have
extremely wide applicability.
The tendency of particles to pair to J = 0+ leads to a scheme in which this
property is explicitly recognized and exploited. Consider the /" configuration.
We ask what is the smallest value of n that can produce a given / value.
Denoting this value by v, it is clear that there can be no particles coupled in
pairs to J - 0 in the configuration f j . (Otherwise, a f-2 configuration would
have a spin/.) Such a state is then said to have seniority v. From a configuration
y+2 we can make a state of the same spin J by coupling one pair of particles to
7 = 0. This state is also said to have seniority v. Physically, v is simply the
number of unpaired particles in a state of angular momentum J in the
configuration;". The number of paired particles is (n - v) and the number of
such pairs is (n - v)/2. For v = 0, all particles are paired and 7 = 0.
Let us further illustrate this concept with a simple example. Consider the
116 Shell Model and Residual Interactions

(f^)4 configuration. From the m-scheme and the simple formula derived
earlier, Jnm = 4j - 4(3)12 = 8. This state can only be made by maximizing the
alignment of all; = 7/2 angular momenta as allowed by the Pauli principle. The
7 = 8 state therefore has seniority 4; there are no particles coupled in pairs to
/ = 0. On the other hand, /=2,4, and 6 states can be made by first coupling one
pair of particles to/ = 0 and then using the remaining | (7/2)2/) configuration to
produce angular momenta of 2,4, or 6. Such states have seniority v = 2. Finally,
the J = 0 state of the (f7/2)4 configuration obviously has seniority 0, that is, all
particles are coupled in pairs to / = 0. (Note that there may be other J = 0,2,
4,6 states of the (f^)4 configuration, all with v = 4.) What we have shown is that
/ = 0,2,4,6 states of v = 0 or v = 2 can be constructed.
The seniority concept is important for several reasons. First, it leads to
many simple, powerful results under very general conditions. For example,
various interactions and matrix elements can be classified in terms of whether
or not they conserve seniority. As will be seen, they have very different
properties as the number of particles in a shell increases. Secondly, and
perhaps most importantly, it seems that many realistic residual interactions
conserve seniority, so this scheme gives reasonable predictions for actual
nuclei. It is impossible within the scope or philosophy of this book to derive all
the results of the seniority scheme without adding an undesirable complexity.
Such derivations are available in many detailed textbooks on the shell model.
The complexity of these derivations often tends to obscure some of the simple
ideas lying behind them. It is these ideas that we wish to emphasize here. We
will derive or motivate a few crucial results; the others can be obtained by
analogous, though more tedious, manipulations.
Perhaps the most important ingredient in understanding the results of the
seniority scheme is the following: consider the / 2 configuration and the matrix
element of any odd tensor interaction. (The introduction of the concept of
tensors and their rank here should not be intimidating. The spherical harmon-
ics of order k, Y^ simply form the 2k + 1 components of a tensor of rank k. An
example of an odd rank tensor is the magnetic dipole operator. The quad-
rupole operator is an even rank tensor. As commented eariler, the 5-function
interaction is equivalent to an odd-tensor interaction.)
For the case of a one-body odd-tensor operator acting in they 2 configuration

The proof of this is trivial. We recall that in the two-particle configuration only
even J values are allowed. Therefore, J on the left side must be even and, by
conservation of angular momentum, there is no way that 7 = 0 can be coupled
to an even J by an operator carrying odd multipolarity.
Equation 5.4 simply states that all matrix elements of one-body odd-tensor
operators vanish in the y2 configuration. This includes the 7 = 0 case. Odd
tensor operators cannot "break" a/ = 0 coupled pair, nor can they contribute
a diagonal "moment." The significance of this simple equation cannot be
overemphasized.
In many-particle systems, it has three enormously important consequences.
For such configurations, one-body operators are normally expressed in terms
 Multiparticle Configurations 117

of sums over operators acting on each particle. A one-body odd-tensor

operator acting in a f configuration is given by U* = "Z^U/1. Since an odd
tensor operator cannot change any }2J = 0) pair to one with J * 0 (J even), odd
tensor operators must conserve seniority. Equation 5.4 shows that there is no
contribution to U* from such pairs of particles coupled to / = 0. Thus matrix
elements of one-body odd-tensor operators in/1 configurations with seniority
v, can be reduced to those in the f configuration. Moreover, they are inde-
pendent of n (for n > v).
These results follow so trivially from Eq. 5.4 that the preceding comments
essentially constitute a derivation. However, they are so important and basic
that it is worthwhile to go through the arguments more explicitly. Consider a
matrix element such as (j"vJ' "£.=1U* \j"vJ). Since v < n, the left side can be
rewritten in terms of wave functions of the configuration ly^vC/')/2^ = O)/)
and similarly for the right side. For simplicity, we take the particles thus
separated off as the (n-l)th and nth particles. Application of Eq. 5.4 to these
two particles contributes nothing to the overall matrix element, and we can
replace the operator "Z.=1IJ * with "~2Z.=1U* extending over n-2 particles. Since
the matrix element is now independent of the last two particles, we can
integrate over them. Since they are in the same state j2J = 0), this integral is
unity by orthogonality. If (n - v) > 4, we can repeat this procedure for another
pair of particles. We continue this procedure for any even v until we are
dealing with an operator "L.^U* acting on the states |;vJ}and 1/7'). Thus we
obtain

which shows both the reduction of a matrix element in the f configuration to

one in;'" and the independence of n.
The other result, conservation of seniority, is equally obtainable. Suppose
the two wave functions in the above matrix element have different seniorities
v, v' < n. There is some point in the successive reduction (the successive
peeling off of pairs of particles) where an overlap integral over the wave
functions fJ = 0} and \j2J ± 0} appears. Clearly, by orthogonality, this
vanishes. To reiterate, we have the absolutely critical results:
• Odd-tensor single-particle operators conserve seniority in ;'" configura-
tions.
• The matrix elements of odd-tensor single-particle operators in;'" configu-
rations in the seniority scheme can be reduced to ones in the f configura-
tion.
• These matrix elements are independent of n.
These rather abstract results have many practical applications. They imply,
for example, that the magnetic moment of the 7/2" state of an (f7/2)3 configura-
tion is identical to that in the single particle L,I2 configuration: in general,
magnetic moments in odd mass nuclei where the valence particles occupy a
given; orbit should be independent of the (odd) number of valence nucleons.
Similar arguments cannot be applied to even tensor operators like the quad-
rupole operator. It turns out that these operators are not diagonal in the
118 Shell Model and Residual Interactions

seniority scheme, but rather connect states with seniorities v and v ± 2. Using
arguments such as these, it is therefore clear why Ml transitions in even mass
nuclei are rare—they can only connect states of the same seniority—while E2
transitions dominate even in near-closed shell nuclei. Therefore this domi-
nance is not necessarily a demonstration of collectivity, but a reflection of the
seniority structure of low-lying states in ff) configurations.
Thus far in our discussion of seniority, we have considered single-particle
operators representing moments or transitions. Equally important are two-
body interactions, which can be either diagonal or nondiagonal. Both are
important, although we will emphasize the former since they determine the
contribution of residual interactions to level energies. A key example is the 5-
function interaction. Clearly, interactions can be written as products of single-
particle operators. We saw an example of this earlier in discussing multipole
expansions of arbitrary interactions. We now turn to consider the properties of
various interactions in the seniority scheme.
Consider an arbitrary odd-tensor two-body interaction V12. This can be
taken as a product of one-body operators, Stodd fj*f2*. As with one-body
operators, it is extremely useful to be able to relate the two-body interaction
matrix elements of seniority v states in the f configuration (n even) to the
matrix elements in a f configuration. Deriving this desired result is trivial.
Consider the matrix element (a subscripted k labels particles, not rank)

where the sum is over the n-particles, and where a and </ denote any addi-
tional quantum numbers needed. Since the states have seniority v (even),
there are (n - v) particles paired off to / = 0. By the same reasoning that led to
Eq. 5.4, the terms in "I. _, kVa, that act on these particles cannot change their
coupling. All that this part of the sum can do is contribute a diagonal matrix
element of the form Q2J = 01 V.J // = 0). But this is just the lowering of the 0+
energy in a;2 two-particle configuration. We define this energy lowering by V0.
The sum contributes this for each such (J - 0)-coupled pair, of which there are
(n - v)/2. Having thus separated off these particles, we are left with a sum over
v particles of the same interaction. Thus, we obtain,

This interaction matrix element may be either diagonal or nondiagonal (in a),
but it cannot change v since it is of odd tensor character. In either case, it is of
absolutely central importance in nuclear spectroscopy. As with the case of
one-body odd-tensor operators, we have an equation relating matrix elements
of a two-body interaction in they" configuration to those in the/" configuration.
Here, however, these matrix elements are not constant across a shell, but linear
in (n - v)/2, the number of nucleons paired off to J = 0. Such matrix elements
peak at midshell. This feature is sometimes known as the pairing properly.
To understand other important implications of this, let us first consider
 Multiparticle Configurations 119

diagonal matrix elements where a = «'. The second term on the right in Eq. 5.6
is simply the number of pairs of particles coupled to / = 0 multiplied by the
interaction energy, V0, for each pair. Recalling that we are dealing with
attractive residual interactions (larger matrix elements imply lower-lying
states), then states with lower seniority v will lie lower in energy. The v = 0
states, which must have /* = 0+, will lie lowest. Immediately, this accounts for
the well-known empirical property that the ground states of (spherical)
even-even nuclei all have /* = 0+.
Similarly for odd mass spherical nuclei, the ground state will usually be a
v - 1, J=j state in which all but one nucleon is paired off in \fj = 0) combina-
tions.
It is worthwhile to explicitly write Eq. 5.6 for a/" configuration in the v = 0,
7 = 0, and v = 1, J = j states. For both situations the first term vanishes since
there cannot be a two-body interaction in a ;v=0 (no particle) or ;'v=1 (one
particle) system. Therefore, the energies are given by the second term:

These equations simply state that the ground state energies in the respective
systems depend solely on the numbers of pairs of particles coupled to J = 0. In
the odd particle case, the unpaired nucleon is, from this point of view, just a
spectator. Indeed, as de Shalit and Feshbach emphasize, the nuclear force
effectively measures the number of pairs of particles coupled to J = 0, at least
insofar as it can be approximated by odd tensor interactions.
One of the most crucial uses of Eqs. 5.6 and 5.7 concerns the energies of
seniority v = 2 states (the following argument applies to higher seniority states
as well, but these are less often identified experimentally). Let us consider the
energy difference E(J"v = 2,7) - E(j"v - 0, / = 0). Simplifying the notation by
denoting the interaction by V, Eq. 5.6 and 5.7 give

Therefore, the energies of the v = 2 states are independent of n. Let us also

calculate the spacings within the v = 2 configuration. These are given by
120 Shell Model and Residual Interactions

Fig. 5.2. Illustration of the constancy of seniority v = 2 levels in/1 configurations.

Thus, all energy differences of seniority v = 0 and v = 2 states in the n-particle

configuration are identical to those in the two-particle system and are inde-
pendent of n. This is illustrated in Fig. 5.2. This result is crucial because its
absence would make it virtually impossible to apply the shell model in a simple
way to nuclei other than those within one or two nucleons of closed shells.
Indeed, this result was anticipated in the previous chapter in arguing that the
shell model has broad applicability. The low-lying levels of good seniority in a
f configuration are independent of n. In practice, more than one orbit; will be
occupied by the valence nucleons. Nevertheless, the present result can be
approximately generalized if one writes the wave function in the schematic
form

In fact, the incorporation of such two-body configuration mixing is essentially

equivalent to a modification of the interaction itself, and thus Eq. 5.6 is widely
applicable. The Sn nuclei (see Fig. 2.6) provide a classic example of Eq. 5.10
and its generalization to the mulli-/ case: the entire known set of v = 2 levels,
J = 0+, 2+, 4+, 6+, is virtually constant across an entire major shell. The Ca
isotopes (Fig. 2.3) provide another example that will be discussed later.
To recapitulate, for the matrix elements of odd-tensor operators and inter-
actions between states of good seniority in/ 1 configurations:
• One-body matrix elements (e.g., dipole moments) are

• Two-body interactions arc linear in the number of

 Multiparticle Configurations 121

It is possible to derive analogous results for even-tensor operators and

interactions. The derivations involve the manipulation of, and recursion
relations for, CFP coefficients. These are tedious but straightforward. We will
cite two important results. Once again, we look first at one-body operators.
As noted, even-tensor operators do not necessarily conserve seniority and can
link states with Av = ±2. The expression for such matrix elements is given by:

Once again, the power of the seniority scheme allows us to link matrix
helements in the configuration/" to those in the configuration;". The square of
Eq. 5.13 gives the behavior of the transition rates induced by the operator
throughout a shell. For large/ and n (/', n » v), this transition probability goes
as (f(l - /)) where / = nl(2j + 1) is the fractional filling of the shell. This
expression at first increases as/, then flattens out, peaking at midshell. More-
over, it is clearly symmetric about the midshell point. Probably the most
common and important application of this concerns E2 transition rates in-
duced by the operator Q = r2Y2. The important quantity (2 + I Q I IO^) 2 is
proportional to the E2 transition rate from the first 2+ state to the ground state
in an even-even nucleus, and can be written for the /" configuration as
[assuming the 2 +(0 +) state has v = 2(v = 0)]:

For shells that are not too filled, so that (2/ ± 1) » n, this becomes

That is, in the/" configuration, the B(E2) value, defined as

is just proportional to the number of particles n in the shell, for small n. For
large n,n-> 2j + 1, it falls off, vanishing, as it must, at the closed shell. For
/, n » 2, we see that, as given in the general case above,

This behavior is commonly observed in real nuclei, with B(E2:21+ -> 0^)
values rising to midshell and falling thereafter. Data beautifully illustrating
this are shown for the Z = 50 to 82, N = 82 to 126 region in Fig. 5.3. (The peak
regions of the B(E2) values in Fig. 2.16 are additional examples of this in
condensed form.) In part, this behavior is due to coherent effects involving
single-particle configuration mixing of different/ values in the wave functions
122 Shell Model and Residual Interactions

Fig. 5.3. Saturation of empirical B(E2) values in the rare earth region that illustrates Eqs. 5.13 and
5.16. The numbers on each line give the neutron number.

for each particle, but the overall behavior still reflects a generalization of this
simple result for the seniority scheme.
For transitions induced by even-tensor operators of rank k > 0 that do not
change seniority, the expression corresponding to Eq. 5.13 is

This equation again expresses an n-particle matrix element for states of

seniority v in terms of the v-particle matrix element. It has an interesting
behavior as a function of n, as given by the factor outside the matrix element.
In terms of/(the fractional filling of the shell), the numerator behaves simply
as (1 - If). It therefore has opposite signs in the first and second halves of the
shell and hence must vanish identically at midshell. This is, of course, an
extremely important result, indicating that, for example, quadrupole moments
of;'" configurations in even-even nuclei change sign in midshell. The generali-
 Multiparticle Configurations 123

zation to many-/ shells suggests that such moments should have opposite signs
at the beginning and end of major shells. Although the trends in realistic cases
are complicated by the different; shell degeneracies, this qualitative feature is
a well-known empirical characteristic of heavy nuclei, and it contrasts mark-
edly with that for odd-tensor operators that are independent of n (see Eq. 5.5).
We shall see another important application of Eq. 5.17 in our discussion of
the p-n interaction in Chapter 6. Moreover, the decrease toward midshell and
symmetry about that point will see important reflections even in deformed
nuclei (systematics of $ Fig. 6.11, p. 164) where seniority is strongly broken.
Finally, we turn to two-body interactions for even-tensor operators. Some
of these interactions can change seniority, connecting states with v and v -2.
For this case, the result is trivial to derive. An even-tensor two-body interac-
tion connecting states with seniorities v and v - 2 must be a product of two one-
body operators—one that conserves seniority, another that connects v and
v-2. In the reduction to a matrix element in the/" configuration, the first gives
a factor identical to that in Eq. 5.17, the second gives the factor in Eq. 5.13.
Thus, their product yields the result

Once again, we note the factor (2/ +1 - 2n), which vanishes at midshell and has
opposite signs in the first and second halves. For;', n » v, this interaction
energy varies across a shell as (1 - 2/)/(l -/): at first this increases with/but it
peaks well before the shell is one-quarter filled, tapers off, and crosses zero at
midshell; in the second half of the shell, it is symmetric to the first half except
for a change in sign.
We will not give the general expression for seniority-conserving matrix
elements in the f configuration since they are more complex, involving not
only matrix elements of the interaction in the f but in the ;v+2 configuration as
well.
It is useful at this point to summarize some of these important results. This
is done in Fig. 5.4, which shows the behavior of both seniority conserving and
nonconserving matrix elements for one-body operators and two-body interac-
tions across a/ shell under the assumption (where applicable) that; and n are
large and much greater than v. For the v -> v - 2 even-tensor case the square
of the matrix element is given since it is directly proportional to the most
common example of such behavior, B(E2: 2^ -> O/) values. Each panel also
gives the (sometimes approximate) analytic formula.
To recapitulate: one-body odd-tensor operators (e.g., magnetic moments)
conserve seniority and are constant: one-body even-tensor operators may
change seniority, with v —»v - 2 transition matrix elements (e.g., B(E2: 2^ —»0, + )
values) peaking at midshell, while seniority conserving matrix elements (e.g.,
quadrupole moments) vanish at midshell and are negatives of each other for
124 Shell Model and Residual Interactions

Fig. 5.4. Summary of the behavior or various operators and interactions across a shell in the
seniority scheme. Note that the middle panel gives the square of the matrix element since this
corresponds to the physically interesting case of B(E2) values.

particles and holes; odd-tensor two-body interactions (e.g., 5-interactions)

behave as (n - v)/2 (the number of pairs of nucleons in J = 0 couplings), and
therefore peak at midshell.
Finally, to relate this discussion to an earlier one regarding the behavior of
various matrix elements for particle configurations and the corresponding
hole configurations, we note that the results in Eqs. 5.5 and 5.17 are consistent
with Eq. 4.33, derived for diagonal matrix elements of one-body operators.
 Multiparticle Configurations 125

Thus far, the discussion has focused on the 5-interaction. A very popular
alternative is the surface ^-interaction (SDI), which, as its name implies, acts
only at the nuclear surface. It is equivalent to the angular part of the 5-
interaction and to the assumption that all radial integrals are equal. Though
this is a simplifying assumption, it permits an important generalization of one
of the key preceding results: for degenerate orbits, the SDI conserves seniority
in the multi-;' configuration I/'"1;'"2- • • » v , j } . (In a single;' shell the 8 and SDI
interactions are identical.) The SDI also has off diagonal matrix elements in
multi-/ situations that give rise to mixed wave functions. These matrix ele-
ments are generally larger than for the volume 5-interaction: the reason is
simply that, when the interaction can occur throughout the nuclear volume,
the effect of non-complete overlap of the particle wave functions is larger.

5.4 Some Examples

With this theoretical background, it is interesting to consider an example that
reflects some of these properties of the seniority scheme. Figure 2.3 showed
the energies of the first 2+ state,£21 for the Ca isotopes as well as the
B(E2:2j+ -» 0,+) values. 2040Ca20 is doubly magic. The lowest orbit beyond the
closed shells is lf7/2 and the ground states of the nuclei from 41Ca to 48Ca are
formed by adding nucleons in this orbit successively. The low-lying states in
the even Ca isotopes can then be viewed as (primarily) an (f7/2)" configuration.
There is an interesting theorem, which we shall not prove, that states that
any two-body interaction in the f configuration is diagonal in the seniority
scheme, provided it is diagonal in the f configuration, that is, if there are no
finite matrix elements connecting v = 3 with v = 1 states. Clearly, since two-
body interactions only connect states of the same J (only states of equal J can
mix), this condition is automatically satisfied for any / value that is not
common to both the fv = 3) and ifv = 1} states. This is useful because it
happens for allj values; < 7/2. It is easy to prove this. For; = 1/2 and 3/2 it is
trivial: they have no v = 3 states since they become maximally filled (midshell)
at n = 1 and n = 2, respectively.
We saw earlier from the m-scheme that for; = 5/2, the only allowed states in
the (5/2)3 configuration are J = 9/2, 5/2, 3/2. The / = 5/2 state can clearly be
formed by coupling a; = 5/2 particle to a | (/ = 5/2)2/ = 0) configuration and so
has v = 1, while the J - 3/2 and 9/2 states must have v = 3.
For; = 7/2 there is, again, only one state with J = 7/2. Of course, it has v = 1
and, since there are no v = 3 states with the same J value, the above mentioned
matrix elements must vanish. So the preceding theorem is trivially satisfied for
all; < 7/2, giving us the useful result that any two-body interaction is diagonal
in the ;'" scheme for; <7/2.
Note the importance of this result: since the first shell model orbit with
;' > 9/2 is the Ig9;2 orbit at the upper end of the 28-50 shell (see Fig. 3.2), this
means that seniority is generally a good quantum number for;'" configurations
for all nuclei with A < 80! These just happen to be the nuclei where shell model
calculations are most feasible (the spaces to be diagonalized are not yet too
large). Of course, this theorem does not apply to mixed-;' configurations
} i l } 2 2 j ] , but it certainly shows why the seniority scheme is so important and
126 Shell Model and Residual Interactions

why seniority is often a reasonably good quantum number even in rather com-
plicated configurations.
In any case, the theorem implies that seniority must be a good quantum
number for the (f^)" states in the Ca isotopes, independent of the interaction.
This does not mean that the interaction energies are independent of the
interaction since, as we saw, the matrix elements of these interactions depend
on their odd- or even-tensor character. Nevertheless, knowing that seniority
must be a good quantum number, we can confidently choose some reasonable
interaction and inspect the predictions of the seniority scheme. We assume
that an odd-tensor interaction is a good choice since the 5-function can be
written in this form. Then, from Eq. 5.6 or, more explicitly, from Eqs. 5.9 and
5.10, the excitation energies of the / = 2+, 4+, and 6+ levels should be constant
across the Ca isotopes.
Let us inspect the data in Fig. 2.3. The yrast energies are high at 40Ca and
48
Ca since both of these nuclei are doubly magic. However, "^.^Ca have, as
expected, much lower 2+l energies and these are indeed relatively constant.
The B(E2) values in the Ca isotopes are also roughly consistent with the
seniority picture since they peak near midshell, although the symmetry about
midshell is not particularly evident.
The constancy of v = 2 level energies in singly magic nuclei has long been
emphasized by Talmi. His classic example is the Sn isotopes, whose energies
were shown in Fig. 2.6.
One can also use the seniority scheme to look at nuclear binding energies,
B.E.^), (n even) for a series of nuclei in which an orbit / is filling. These
binding energies are the absolute energies of the I f , v - 0, J = 0+> ground states
(we assume an odd-tensor interaction). Of course, this is only the residual
interaction energy, to which must be added the n single-particle energies Enl..
To be slightly more general, there is one other interaction that is diagonal in
the seniority scheme—the trivial case of a scalar interaction. Since this is a
constant, the interaction matrix element will be equal to that constant multi-
plied by the total number of pairs of nucleons that can interact, n(n - l)/2.
Combining these three terms, one then has for the binding energy in a /"
configuration, relative to the closed shell,

This formula, which is valid for any interaction that is diagonal in the seniority
scheme (and therefore, for any interaction for; < 7/2) displays the well-known
parabolic behavior of nuclear masses. We note that the parameter B always
turns out to be negative: the quadratic term is always repulsive. Evidence of
this can be seen empirically from the data presented in Chapter 1, which
showed that binding energies are approximately proportional to A. If B > 0
(attractive), binding energies would increase quadratically, instead of linearly,
with A.
This is the same point we have made before, that the nonpairing, residual
interaction between like nucleons is repulsive. We can now carry this one step
further. Since the quadratic term stems from the scalar part of the interaction,
 Multiparticle Configurations 127

which is obviously long range, it is the long-range component of the interaction

between like nucleons that must be repulsive.
This conclusion, based on the simple form of the binding energies in the
seniority scheme and the empirical behavior of separation energies, forces us
to conclude that deformation does not arise simply from an abundance of
valence nucleons outside closed shells, but must specifically involve both
valence protons and neutrons. We will see later how the properties of the
proton-neutron interaction can indeed lead to deformation through the ef-
fects of one-body configuration mixing.
Most of the examples of the seniority scheme so far in this chapter have
concerned even nuclei. However, the scheme is equally powerful in treating
odd mass nuclei near closed shells. We will illustrate this with a simple
calculation that is useful in considering low-lying energy levels in sequences of
odd A nuclei.
As we pointed out, the lowest state of the f configuration in an odd mass
nucleus normally has / = / and v = 1. The n-dependence of its interaction
energy is [(n -1)12] V0 where VQ is the interaction energy in the / = 0 state of the
j2 configuration. For an odd-tensor interaction, the excitation energies of the
v = 3 states can be obtained from Eqs. 5.6 and 5.8:

Thus, these excitation energies are identical to those in the/ 3 configuration and
are independent of n—that is, they are constant across a shell. Clearly this also
means that the spacings between v = 3 states are n-independent. This can of
course be seen by explicit calculation:

In closing this discussion of multiparticle configurations, it is interesting to

take a more complicated level scheme as an example and see how far we can go
in interpreting it by exploiting the simple results in this chapter. Consider then
the nucleus ^Nd^ with ten protons outside doubly magic132Sn (see Fig. 5.5).
Although the level scheme seems rather complex, nearly every feature can be
easily understood and, indeed, derived analytically, without any complex
calculations.
128 Shell Model and Residual Interactions

Fig. 5.5. Low-lying levels of 142Nd in comparison with analytic and numerical shell calculations.
(The empirical level scheme is based mainly on Wirowski, 1988.)

To start, we note that the lowest shell model orbits in the Z = 50-82 shell are
2d5/2 and Ig^. There are many ways that the protons can be distributed over
the two positive parity orbits. If we assume that the splitting of states of a given
seniority v is small compared to the spacings between states of different
seniorities (we are neglecting the first term relative to the second in Eq. 5.6),
then the v = 0 / = 0+ state will, as usual, lie lowest, by an amount V0 below the
v = 2 states. The v = 2 states, in turn, occur, on average, this same distance
below the v = 4 states. Therefore, we can assume that all of the J * 0 low-lying
positive-parity states (J* < 6+) are v = 2. Since we are dealing with (positive
parity) j values < 7/2 we know that seniority is a good quantum number
regardless of the interaction. Since I (ds/2)" v = 2} does not give a 6+ state, let us
assume for simplicity that the ds/2 shell is filled, leaving four protons in the g7Q
orbit. By Eq. 5.6 and the preceding discussion, the J = 2+, 4+, 6+ v = 2 excitation
energies should be identical to those in a (g7/2)2 configuration, and as such can
be estimated theoretically using the results for a 5-function interaction. We
postpone doing so for a moment until we determine how best to estimate the
absolute strength of the interaction. The next positive parity state is 8+ and
must have v = 4. By our earlier arguments it should lie higher than the v - 2
states by roughly \V0\. The lowest-energy way to construct it is to couple the
two v = 2 states with/^ 6+ and 2+ together. Similarly, the 10f level must also
have v = 4. The easiest way to form it is by coupling the J" = 6+(v = 2) and
j" = 4'(v = 2) states.
 Multiparticle Configurations 129

The negative parity states must involve the hn/2 orbit. The two lowest con-
figurations should involve eight protons coupled to / = 0 in the d5/2 and g7/2
orbits and an (h^ g7/2) or (hn/2ds/2) pair. The former gives spins/ = 2~ - 9~ while
the latter yields / = 3~ - 8\ From the rules developed in Chapter 4 for the
ordering of different / states in two-particle configurations under the influence
of a 5-interaction, we find that the 9" state should be the lowest in the
| hjj^gj^/} configuration while the 3~ should lie lowest in the I hn/2d5/27) con-
figuration. In both cases the even-spin negative-parity states are unaffected by
the interaction. Empirically, the lowest negative parity state is indeed 3~, sug-
gesting the (h^d^) assignment. The lowest T and 8" levels must also belong
to this multiplet (since, if they were part of the (h^g^X/ = 2~ - 9~) multiplet
they would lie above the 9~). The 8~ level gives the unperturbed position of the
(hu/2d5/2) multiplet, and the energy difference 8"- 3~ gives the absolute scale of
the interaction strength, thus allowing us to predict (using Table 4.1) the 5" and
T energies, as well as the spacings among the positive parity levels. The J = 9~
state is then the lowest member of the (hn/2g7/2) multiplet. (The lower-spin
members would not have been detected in experiments carried out for 142Nd.)
Finally, the J* = 10~ and 11- states cannot arise from either (hn/2g7/2) or
(hn/2d5/2) couplings. Since, empirically, the latter lies lower, a reasonable con-
figuration for the 10" level is I (hn/2d5/2)/ = 8~ ® 2+, )10~, meaning an (hn/2d5/2)
pair coupled to /* = 8" built on the seniority v = 2 2+ level of the remaining (n-2)
particle system. The energy difference £(10~) - £(8") ~ £ 2 J is approximately
satisfied experimentally. The lowest 11~ levels can be made either by coupling
the h1]/2g7/27 = 9")state to the v = 22+l level or the hn/2d5/2./ = 7~) level to the
v = 24\ level.
All these results are incorporated now into Fig. 5.5, where it is evident that
the agreement of this extremely simple calculation with experiment is actually
remarkably good. The ordering and energies of most of the levels are correctly
predicted analytically, in agreement with experiment, and with only three
parameters: the single-particle energy differences ehn( 2 -e g 7/ 2 and ed 5 | 2 -e g7 / 2
and the strength of the residual interaction. Figure 5.5 also shows an actual
detailed 10-particle diagonalization of the 142Nd level scheme, using a surface
5-interaction (which gives relative spacings, within a configuration Iff), which
are the same as for a volume 5-interaction and which is otherwise similar as
well, although different in some details) with strength 0.4 MeV and with the
single-particle energies (in MeV): eg7|2 =0, EdS|2 =0.7, £ im/ 2 = 2.5. The
calculation also shows reasonable agreement with the empirical scheme, but
more importantly, it shows that our simple analytic interpretation is a re-
markably accurate approximation of a complex shell model diagonalization.
This kind of interpretation highlights the power of the methods we have
discussed and shows how far we can go in a relatively simple shell model
interpretation of rather complex level schemes. The principle difference in the
numerical diagonalization is that additional components, such as (d5;2)2 and
(d^^) or even (ds/2g37/2), come into play. Similar analyses can be applied to
countless other nuclei (e.g., Sn) and greatly help to understand the results of
complex realistic calculations.
Improved calculations, compared to Fig. 5.5, have also been carried out.
130 Shell Model and Residual Interactions

They allow (hn/2)2 as well as (ds/2)nl and (g7/2)"2 configurations in the positive
parity levels. The inclusion of these amplitudes mainly affects the required
hj^ single particle energy and, of course, changes the strength of the interac-
tion needed to fit the data. The reason is easy to see from Eq. 4.10. Since the
lowering of the 0+ state in a configuration \j 2J) is proportional to (2/ + l)/2
multiplied by the interaction strength, V0, the inclusion of (h11/2)2 means that a
smaller \V0\ is required to maintain the same / rast- /0+ spacing. The extra
lowering of the 0+j level effectively raises the excitation energy of all the
others. To regain a fit to the data for the negative parity states, a lower ftn/2 is
required.
This illustrates two important points. First, the choice of effective residual
interaction, single-particle energies and the shell model space are intimately
linked. One should be wary of conclusions regarding any of these if there is not
supporting evidence for the choices concerning the others. Secondly, despite
the beauty of analyticity, a realistic treatment of complex nuclei still requires
detailed explicit calculations if really quantitative results are desired.
Finally, note that we have completely ignored core excitations of the pro-
tons or neutrons. These can be significant and their effects can vary signifi-
cantly for different states. In particular, even in singly magic nuclei (e.g., 142Nd
or Sn), the lowest 2+ and 3" states are often rather "collective" with a number
of major components, including core excited particle-hole components (e.g.,
(g^-1 hn/2) / = 3~ for neutrons in Nd or protons in Sn). To some average extent,
their ignored effects are mocked up by the choices of the single-particle
energies of the valence orbits and of the interaction strength. It is no wonder
that residual interactions are often called "effective interactions" and that an
extensive theory of such interactions has been built up. Indeed, as discussed at
the end of Chapter 4, a number of alternates to the <5-function interaction are
often used. Often, these are more structured, finite-range, interactions such as
Gaussian forms. Various so-called Skyrme interactions are also popular, as are
interactions defined explicitly in terms of sets of two-body empirical matrix
elements. It is clear from the discussion at the end of Chapter 4 that such finite-
range interactions have different (often more parabolic) ./-dependencies than
the 5-function.

5.5 Pairing Correlations

In closing this chapter, we want to turn to another interaction, similar in many
ways to the 5-interaction, which is very important in understanding the struc-
ture of heavy nuclei. We saw in Chapter 1 that proton and neutron separation
energies exhibited an odd-even effect, indicating an extra binding in the 0+
ground states of even-even nuclei. We have also seen that the 5-function force
acting on two identical nucleons in equivalent orbits produces a strong lower-
ing of the / = 0+ state. This behavior is typical of any very short-range
interaction and is a direct consequence of the Pauli principle, which allows
such interactions to affect only spatially symmetric wave functions. Although
the 5-function force is particularly easy to deal with (and has specific tensor
properties that imply conservation of seniority), it has become traditional in
 Multiparticle Configurations 131

heavy nuclei to describe such behavior more directly in terms of the so-called
pairing interaction. This is defined (see Eq. 4.20) to be an attractive interaction
acting only on two identical particles in total angular momentum 0+ states.
States with /* * 0+ are unaffected. A comparison of the pairing interaction with
a 5-function interaction for an (f7/2)2 configuration was shown in Fig. 4.13. The
sequence of energy levels for the 5-function interaction is the familiar result we
have been studying and simply reflects the properties of the 3 -/ symbol in Eq.
4.10. The pairing force presents a similar overall pattern but with a degenerate
multiplet of J* * 0* states at the unperturbed energy. As we have seen,
inspection of the empirical level schemes of two-particle states in singly magic
nuclei such as 210Po, 210Pb, 134Te (Fig. 4.5), and many others, shows that the 5-
interaction reproduces empirical spectra much more accurately. Even for
multiparticle (but still singly magic) cases, such as the Sn isotopes, the 5-
function provides a more realistic interpretation (see Fig. 2.6). Nonetheless, it
has become nearly universal practice to invoke a pairing interaction and to
refer to the lowering of the 0+ state as a pairing effect. (Indeed, we already
referred to the second term in Eq. 5.6 as the "pairing property" in analogy with
the effect of the pairing force.) Part of the appeal of the pairing force is the ease
with which it can be extended to multiparticle systems, where it leads to the
desired result of 0+ ground states in even-even nuclei without the need for the
seniority apparatus and angular momentum algebra we have discussed.
There are a number of experimental facts that motivate the introduction of
a pairing force and the concept of pairing correlations. We have mentioned
some, but it is useful to summarize them here. The best known is the simple
fact that the ground state of all even-even nuclei has J* = 0+. A related point is
that this 0+ state is normally far below other noncollective intrinsic states. This
is the so-called pairing gap.
Perhaps the most direct evidence for a pairing interaction is the so-called
odd-even mass difference. This simply refers to the fact that when nucleons
are added to a nucleus, the gain in binding energy is greater when an even-even
nucleus is formed than when the neighboring odd mass nucleus is formed. This
empirical fact can be inferred from the data of Figs. 1.2 and 1.3 by comparing
the absolute values of S(n) or S(p) for adjacent odd and even nuclei. An extra
attractive interaction that couples pairs of like-nucleons together can accom-
modate this fact, and indeed, the separation energy data suggest a strength for
the pairing interaction of -1-2 MeV.
There are three other features that can be seen, at least in retrospect, as
clearly pointing to the need for a pairing interaction. The pairing interaction
clearly favors sphericity since it favors the formation of pairs of particles
coupled to a total magnetic substate M = 0. Therefore, near closed shells, the
presence of a strong pairing interaction will inhibit the tendency to deform.
Instead of the smooth transition toward deformation that would normally
occur as valence nucleons are added, one typically sees, empirically, a se-
quence of more or less spherical nuclei followed by a rather rapid transition
region to deformation. Secondly, for a deformed nucleus of a given shape, it is
easy to calculate the moment of inertia. This in turn determines the spacing of
132 Shell Model and Residual Interactions

rotational levels (see Chapter 6). Empirical moments of inertia extracted from
those energies are systematically lower than calculated in the shell model. The
inclusion of pairing correlations essentially removes the discrepancy. Finally,
the energy distribution of Nilsson orbits near the ground state in deformed odd
mass nuclei is, on average, denser than expected from the Nilsson diagram.
This will be seen to be a "compression" effect that occurs in the transformation
from particle-hole to quasi-particle energies.
The pairing interaction was defined in Chapter 4 in terms of its matrix
elements as:

where G is the so-called strength of the pairing force and the rest of the
expression has an obvious meaning. Note that the pairing force is independent
of orbit but, since it is identical for each magnetic substate, scales for an orbit
/ as (2/ + 1). It is therefore stronger in high j orbits. Although G is orbit
independent, it decreases with A in heavier nuclei where the outer nucleons
are generally further apart and so spatial overlaps are likely to be less. G may
be different for protons and neutrons, being lower for the former because of
Coulomb repulsion. Commonly used prescriptions are

It is frequently imagined that the pairing force is an interaction only

between two particles in the same; state, coupled with their angular momenta
antiparallel, to form a /*= 0+ state. This, of course, is an important facet of the
pairing interaction. However, there is another vital ingredient that is evident
from its definition. The pairing force is equally strong for matrix elements
connecting a I f - J - f r ) state with a i/' 2 / = 0+) state. That is, the pairing force
has nondiagonal as well as diagonal components that can "scatter" pairs of
particles from one orbit to another. Note the importance of this point. If the
force were purely diagonal, then, while 7=0* pairs would still be tightly bound,
excited states could be formed simply by raising both particles in such a pair to
the next unoccupied orbit. On average, this would require twice the energy
needed to raise a single particle from orbit/to orbit/' in the absence of pairing.
The "energy gap" in even-even nuclei would then be only twice the average
spacing of low-lying levels in the adjacent odd nucleus and not 5 to 10 times as
large, as observed experimentally. Rather, by scattering particles from one /
orbit to another, the pairing force mixes 0+ states and creates partial occupan-
cies near the Fermi surface. Hence it builds up a coherence in the pair wave
functions, which further lowers the lowest 0+ state and thereby enlarges the
gap: if we picture the pairing force in a perturbation theory context, the
amplitudes for scattering a pair of particles from orbit / to orbit / will be
proportional to the matrix element (f-J = 0\ VfJj'2J = 0 ) l(e. - e ) =Gl(ej - e..).
In the absence of pairing, all levels would be occupied up to some point (the
Fermi energy) while those above this energy would be completely empty. With
pairing, however, many orbits can be partially occupied. This in turn, radically
alters the concept of hole excitations, and thereby, the levels schemes of both
odd and even nuclei. We shall see all this a bit more formally when we outline
 Multiparticle Configurations 133

Fig. 5.6. Definition of several quantities used in the calculation of pairing correlations. (Left)
Idealized set of equally spaced single-particle shell model levels. (Right) Resultant orbit
occupancies V2. The calculation is general since the single-particle energies are given in units of A.

the basic results of pairing theory. We will not derive these standard results but
will try to highlight their key effects.
To proceed, we refer to Fig. 5.6 where several quantities important in
discussing the pairing interaction are defined: the Fermi energy, denoted A;
the single-particle energies, e, with ea being reserved for that level closest to
the Fermi surface; and A, the so-called gap parameter, defined in terms of a
hsum over orbits i, j as

where the usual U and V factors are the so-called emptiness and fullness
factors that pervade the study of heavy nuclei. They are given by:
134 Shell Model and Residual Interactions

These equations can be solved for A, U., and V. for a given set of single-particle
energies e. Or, A can be estimated from empirical mass differences between
adjacent nuclei with odd and even numbers of nucleons. The behavior of V?
against the ratio (e. - A)/A is shown in Fig. 5.6. We see that Vi -> 1 for
(e. - A) « 0 and vanishes for levels far above the Fermi surface. The opposite
applies to Uf. Both fall off rapidly for single-particle energies e. ~ A. Also, from
Eqs. 5.23, V2 + U2 = 1. Uf is the probability that the orbit i is empty, whereas
V.2 is the probability that it is filled. Pairing smooths out the level occupancies
near the Fermi surface over a range ~A, which, in turn, is proportional to G, the
"strength" of the interaction. If there were no pairing, the Fermi surface A
would coincide with the last filled orbit being filled and (e. - A.) would be the
excitation energy required to excite one of the nucleons in this last orbit to one
of the higher orbits er In the presence of pairing, however, this single-particle
excitation energy (e. - A) is replaced by a quasi-partide energy E. given by

Thus, particles and holes are replaced by quasi-particles representing partially

filled levels, and a particle-hole excitation is replaced by the creation of one
quasi-particle and the destruction of another.
The transformation from particles to quasi-particles is of interest not only
because the pairing interaction happens to lead to partial occupancy, but
because it allows an enormous simplification in shell model calculations with
many valence nucleons. It is basically a transformation from a viewpoint based
on the closed shell to one based on the Fermi surface. Thus, instead of having
to deal with all possible ways of constructing each J state within a major shell,
it naturally produces a physically intuitive truncation scheme directly keyed to
the scale of excitation energies one is interested in. For low-lying states, one
needs to consider only low energy quasi-particles. In the shell model, in
contrast, a near-midshell nucleus would involve single-particle energies lying
rather high in the shell regardless of the excitation energy. Moreover, pair
scattering would give rise to extremely complex wave functions in the shell
model. In the quasi-particle picture, all the complexity due to partial pair
occupancies induced by the pairing force is effectively absorbed into the
ground state (quasi-particle vacuum) and we need only consider quasi-particle
excitations relative to the Fermi surface.
The behavior of Ei is interesting and has important consequences. If e. - A,
that is, if the z'th single-particle level is near the energy where the occupation
probabilities fall off rapidly, E. ~ A. This value is also clearly the minimum
value of E..
In an odd mass nucleus, however, this is not the minimum energy for an
excited state: Rather, excited levels are obtained by replacing the quasi-
particle defining the ground state by one corresponding to a different single-
particle level. Thus the excitation energy E°t for the ith quasi-particle in an
odd mass nucleus is given by
 Multiparticle Configurations 135

Fig. 5.7. Effects of pairing in odd and even nuclei. (Left) Level compression near the ground state
for odd mass nuclei. (Right) The energy gap in even-even nuclei.

where EQ(eQ) is the quasi-particle (single-particle) energy of that orbit nearest

the Fermi energy A. Thus we see that £*, can take on arbitrarily small values.
Indeed, since all E. ~ A for (e. - A) « A, the effect of pairing is actually to
decrease the excitation energies of the low-lying states, compressing the excita-
tion energy spectrum. Figure 5.7 shows this compression of E°f as a function
of (e. - A) in units of A. At higher energies [(e - A)»A], the effect is to lower
all states by an amount ~A since Et —»(e. - A) and
In even nuclei, the effect of pairing is nearly the opposite: instead of a
compression of excited quasi-particle levels, there are no simple excitations
below £/ = 2A. This is easy to show. The simplest excitation consists of
breaking one pair and raising a particle to the next higher orbit. Without
pairing, this is a particle-hole excitation. In the presence of pairing, it appears
as a two quasi-particle excitation, one quasi-particle being the hole left behind
and the other being the particle excitation newly created. Thus, the excitation
energy is given by
136 Shell Model and Residual Interactions

as well. It is immediately clear that the minimum energy is Ej = 2A, giving the
famous "pairing gap" that is a nearly universal feature of even-even nuclei in
which few two quasi-particle excitations appear below -1.5 - 2 MeV. In fact,
this empirical energy gap is one way of extracting A experimentally. As noted,
typical values of A range from about 700 keV -1 MeV.
Although there is a large gap between the ground state of an even-even
nucleus and the first excited two-quasi-particle state (2A), above this energy
the quasi-particle level density will be at least as large as in a neighboring odd
mass nucleus.
The energy gap and the high-level density just above it are evident in
virtually any even-even nucleus. Figures 2.5, 2.6, and 2.10 show several
examples. However, there is another very simple and elegant way of demon-
strating the pairing gap empirically without the need for in depth study of each
individual excited state. The («, A) reaction proceeds by the capture of a
neutron of energy En into a target nucleus, forming a residual nucleus at an en-
ergy Ec - S(n) + En. This "capture" state then decays by the emission of yrays
to low-lying levels. Each y-ray thereby defines an excited state by the equation
Exl = Ec - Ey. Under appropriate experimental conditions (See Chapter 10)
the (n, y) reaction provides an a priori guarantee of observing all final states in
certain spin and excitation energy regions. Thus, a spectrum of y-ray transi-
tions will reveal a sequence of peaks, each corresponding to a specific excited
level, and together, displaying all the excited states with certain /* values.
Because of the "completeness," the pairing gap should be immediately identi-
fiable by a sudden increase in peak density. Figure 5.8 shows such a spectrum
for the case of 196Pt, and includes an insert portraying the reaction process
schematically. One sees the ground state, a few low-lying excited states, and

Kg. 5.8. Spectrum of primary frays following average resonance capture (ARC) into 196Pt. The
inset schematically illustrates the reaction process. Two peaks in the spectrum are labeled by their
y-ray energies in keV: the rightmost peak corresponds to a transition to the ground state, while
that at 6100 keV populates a state at 1823 keV The increased level density at the pairing gap is
immediately obvious at approximately this excitation energy (Cizcwski, 1979).
 Multiparticle Configurations 137

then, at Ef -1.8 MeV, a dramatic increase in level density. By inspection, one

can immediately point to an approximate value for 2A.
We also note that, seemingly contrary to the preceding discussion, there are,
in fact, a few peaks below the pairing gap. We shall discuss these in consider-
able detail in later chapters. They correspond to macroscopic "collective" vi-
brations and/or rotations of the nucleus as a whole and can be viewed as
resulting from the interactions among a number of two-quasi-particle states in
which the lowest such state is pushed down by mixing, and therefore occurs
below the energy gap.
Finally, the defining equation for A is interesting since it indicates something
about the origin of the pairing gap. Clearly A should vanish in the absence of
pairing, which the proportionality to G in the Eq. 5.22 reflects. Secondly, the
range and number of energy levels over which U and V are different from 1 and
0 increases with G, as does the sum in Eq. 5.22.
To summarize the discussion of pairing, it is an extremely important con-
tribution to the structure of medium and heavy nuclei, affecting energy levels,
y-ray transition rates, particle transfer cross sections, moments of inertia, and
the microscopic structure of virtually all collective excitations. It modifies
most nuclear matrix elements. It is also, as we commented in Chapter 4, one of
the two ingredients in the so-called pairing-plus quadrupole (or PPQ) interac-
tion, which is a standard basis for the study of collective excitations (See
Chapter 9). We shall see the effects of pairing throughout the rest of this book.
To conclude this chapter and lead into the next part of the book we
emphasize that, despite the numerous successes of the shell model over the
years, it is apparent that it is not actually applicable for the majority of nuclei.
Even the complex M2Nd case just discussed required simplification by ignoring
the excitation of the closed proton and neutron shells. When one deals with
valence nucleons of both types, the shell model rapidly becomes intractable,
especially in the larger multi-/valence shells of heavy nuclei. Moreover, as we
shall see, such nuclei display characteristics that clearly point to nonspherical
or deformed shapes. Thus, there are better discussed in the context of
geometric (and other) models where such shapes arise naturally. We now turn
to a discussion of such models. Later we will show how the collective features
that characterize these models can, in fact, be obtained (and therefore justi-
fied) microscopically from the shell model.
This page intentionally left blank
 PART III

C O L L E C T I V I T Y , PHASE T R A N S I T I O N S ,
DEFORMATION
This page intentionally left blank
 6
COLLECTIVE EXCITATIONS IN E V E N - E V E N
NUCLEI: VIBRATIONAL AND ROTATIONAL
MOTION

6.1 An Introduction to Collectivity, Configuration Mixing and Deformation

The shell model is generally considered the fundamental nuclear model.
Historically, it was the first model to have considerable, detailed success. Of
course, this is not accidental. The shell model works best for light nuclei.
Much of the crucial information on nuclear structure comes from nuclear
reactions that require the incident projectile to have energies near or above
the Coulomb barrier in order to penetrate the nucleus. This requires higher
and higher energies for heavier and heavier elements and, consequently, most
early studies with low-energy accelerators also focused on light nuclei. The
development of nuclear models is intimately connected to the historical prog-
ress of experimental techniques.
More fundamental to the shell model's central position in nuclear physics is
that it provides a well-defined procedure for the calculation of basic nuclear
observables. The principle ingredients needed for any given calculation are a
choice of the single-particle energy levels, of the number of these that should
be included in the space to be diagonalized, and the residual interaction to be
used. Moreover, since the shell model is the only broadly applicable micro-
scopic model available, it is the standard against which others are compared,
and it is the source and rationalization of macroscopic models. A collective
model that can be shown to be inconsistent with the shell model is discarded
with little delay.
Unfortunately, the use of the shell model is, in practice, rather severely
limited. In previous chapters we considered some simple applications, mostly
those in which the particles were confined to a single ; shell. Except for very
light nuclei, however, and those very near closed shells, the valence nucleons
occupy more than one; shell. (The pairing interaction assures this if the single-
particle spacing is less than or comparable to its "spreading" parameter A.)
While shell model calculations for multi-y-configurations are certainly pos-
sible, the size of the matrices in which the residual interaction must be
diagonalized rapidly becomes enormous. For even a few valence nucleons in
several;' orbits, it is easily possible to construct hundreds of states of a given J*
value. Even if such calculations are possible using high speed computers, the
results are difficult to interpret physically and the consequences of agreement
or disagreement with the data are much less intuitively informative. The
situation becomes totally intractable when more valence nucleons are added.
141
142 Collectivity, Phase Transitions, Deformation

To quote the famous example of Talmi, in ^Sm^, approximately 3 x 1014 2+

states can be constructed with protons and neutrons in the Z = 50-82 and N =
82-126 major shells. In such a case, no shell model calculation is even remotely
feasible, nor could one even begin to understand the resultant wave functions.
Since the vast majority of nuclei fall into the many-valence nucleon category,
any advance in our understanding of nuclear structure beyond a very select
group of nuclei must depend either on some truncation of the shell model or on
the development of alternatives to it. This problem is, of course, a common
one in physics. One develops a simple model. Later, one introduces first-order
perturbations to it to refine its predictions. Further improvements involve
incorporating additional degrees of freedom, which enlarge the basis. When
the required basis becomes too large or the wave functions too complex in that
basis, one looks for a new physical picture that allows one to approach the
problem from an alternate basis.
The development of nuclear structure in medium and heavy nuclei has
followed two related paths, which we shall discuss extensively in the next three
chapters. Both involve the concept of a nonspherical shape, but one empha-
sizes the single-particle motion of nucleons in a field of that shape, while the
other stresses the macroscopic motions and excitations of a nucleus having
that shape. Ultimately, the former approach (a "deformed shell model") can
be shown or must be shown to be the microscopic justification for the latter,
but in practical terms, the two models are often used complementarily in a
microscopic-macroscopic combination that has proved to be very powerful.
The nonspherical, or deformed, shell model approach is usually called the
Nilsson model, while the macroscopic one is a fundamentally collective model,
generally known as the collective model of Bohr and Mottelson although
many others have contributed in essential ways to it. Each of these ap-
proaches, of course, has numerous offshoots, extensions, and refinements.
The assumption of a nonspherical macroscopic nuclear shape is a phenom-
enological or ad hoc one: while the enormous successes of the collective
model, its demonstrated predictive power, and the numerous progeny it has
spawned over the last three decades leave little doubt that it aptly describes the
nuclear structure of perhaps the majority of nuclei, one is left with an uneasi-
ness about the apparent incompatibility of an independent particle picture
such as we have been discussing and the clearly collective and coherent motion
involved in macroscopic rotations and vibrations. It was a theoretical achieve-
ment of basic importance in the early 1960s when it was demonstrated that
macroscopic collectivity could indeed result from the shell model with appro-
priate and reasonable realistic residual interactions. The essence of the
method involved is the so-called random phase approximation or RPA method,
or its somewhat simpler cousin the Tamm Dancoff Approximation (TDA).
We shall discuss these two methods in some formal detail in Chapter 9.
In the last decade or so, a rather different approach to collective behavior in
nuclei has been developed. It exploits the dynamical symmetry structure of
nuclei, and utilizes powerful group theoretical techniques to obtain many
nuclear properties by simple algebraic techniques, often in analytic form.
There are many such algebraic models in use today. However, beyond a doubt,
 Collective Excitations in Even-Even Nuclei 143

the most popular, successful, and widely tested to date has been the interacting
boson approximation (IBA) model of lachello and Arima. This model will
also be discussed later in this chapter. Other algebraic models, which empha-
size the fermions (nucleons) directly, currently offer enticing initiatives that
promise to be both more general and more microscopic.
Since the shell model picture of nucleons orbiting in a spherically symmetric
central field must ultimately be relied upon for the microscopic justification of
collective behavior in nonspherical nuclear shapes, it is worthwhile to see how
this model is capable of producing nonspherical configurations. To many
students this is a mysterious point: the explanation is actually quite simple
while the confusion stems from a semantic misunderstanding. This discussion
will form at least a qualitative justification for the basic concepts of the Nilsson
model—the assumption of a deformed shell model potential—as well as
highlight the central importance of the proton-neutron (p-n) interaction in
the development of collectivity and deformation in nuclei. And finally, it will
help to foster an appreciation of the key role played by the distribution of
valence nucleons between protons and neutrons, as opposed to a simple
consideration of the total number of valence nucleons.
There is really nothing mysterious in the idea of generating deformed
shapes from the spherical shell model. The model is spherical in the sense of
the shape of the central potential, not the resulting shape of the nucleus. A
spherical potential allows no distinction between magnetic substates of a given
orbit, all of which are degenerate. A single-particle in a shell model orbit
occupies all such substates equally, and overall, its wave function will be
spherically symmetric. However, any mechanism that yields an unequal
occupation of m states gives a nonspherical shape. (An orbit may be circular,
but it is a circle in a specific plane and therefore gives the nucleus a "bulge" in
that plane.) Such mechanisms abound in the shell model without introducing
anything fancy (even without residual interactions). Consider, for example, a
configuration of two identical nucleons in equivalent orbits, say I (f7/2)2/),
outside a magic core. The 0+ ground state is spherical, of course, but let us look
at the construction of the 2+ state in terms of m substates. From Table 5.1 we
see that many possible mv m2 values, such as m^ = 7/2, m2 = 1/2 or m^ = 5/2,
m2 = 3/2 give M = ml + m2> 2 and cannot contribute to the 2+ state. This state
is then formed from a nonuniform distribution of mv m^ components and must
be nonspherical. Indeed, this is why it has a quadrupole moment (see Eq. 5.5).
A nonuniform magnetic substate distribution is in fact so characteristic of
deformation that one of the best known features of the Nilsson (deformed
shell) model is a filling of orbits based on their m values instead of their;
values. But that jumps ahead of the discussion. We must discuss how and why
/ itself nay not always be a good quantum number. This is an essential point
since "configuration mixing" of single-particle; values ensures an unequal m
substate distribution and is therefore tantamount to ensuring deformation.
To do this we will show that there is a fundamental difference between the
occupation of valence orbits by like nucleons (e.g., two protons or two neu-
trons) and unlike nucleons (one proton and one neutron). Consider then, and
by way of example, a nucleus with two valence nucleons filling the lower part
144 Collectivity, Phase Transitions, Deformation

of the 8-20 shell in the ld5/2 and ldw orbits, which are separated by =5 MeV.
We want to consider matrix elements that can admix d5/2 and d3/2 compo-
nents. In Chapter 4 we discussed diagonal matrix elements of short-range two-
body residual interactions. Now, we are dealing with nondiagonal ones
(although the diagonal elements still play an important role in modifying the
unperturbed energies of the states that mix). Nevertheless, the basic idea is the
same: if the particles are not close to one another in the two two-particle
configurations, the matrix element will be small. The Pauli principle must also
be considered. In addition, as opposed to the diagonal case, here we also need
to consider the unperturbed (initial) energy spacing of the two configurations.
The whole issue then is just one of two-state mixing. We consider the possible
matrix elements:

and concentrate on qualitative effects, ignoring complexities due to angular

momentum coupling coefficients.
For like nudeons (T = 1) the unperturbed energies of the (d5/2)2 and (d3/2)2
configurations are -10 MeV apart. Though the individual / states of each are
lowered by the diagonal residual interaction (see Fig. 4.3 (top)), this lowering
is roughly similar in the two configurations. The spacing thus remains =10
MeV, and it is unlikely that strong mixing will occur. (Recall the pairing
discussion: A ~ 1 MeV and states are admixed only over an energy range of that
magnitude.)
The mixing is also small for matrix elements like {d25/2J I V I d5/2d3/27), since
although the unperturbed separation is now only ~ 5 MeV, the Pauli principle
enters in an important way. The configuration (ds/2)2 for like nucleons only
exists in / = 0, 2,4 states. The (ds/2d3/2) configuration does not exist as / = 0.
Therefore, the strong/ = 0 interaction is forfeited and we are left with only the
J = 2 and 4 cases. But, we can apply our geometric analysis of Chapter 4 to these
cases. For / = 4, for example, the angle 9 between the two orbits in a (d5,2)2
configuration is ~ 82°, while it is = 49° for (d5/2d3/2). We can see the effect of this
if we imagine that one orbit in each configuration is fully aligned with one in
the other (the optimum case). Then, to couple these two states (that is, for a
finite matrix element), the short-range residual interaction must act over an
angular "distance" A0 ~ 33°. We have seen in Chapter 4 that such matrix
elements are small. Thus, in all these cases, the like nucleon configuration
mixing amplitudes will be small.
For unlike nucleons, the large energy difference (=10 MeV) between (d5/2)2
and (d3/2)2 configurations again leads to small mixing. However, the single
nucleon mixing induced by the (ds/2271 V d5/2d3/2/) matrix element is not neces-
sarily small. First, the spacing is only =5 MeV. Second, both configurations
exist in / = 1,2,3, and 4 states and, third, the angles involved favor large matrix
elements. For example, for/ = 1, the angle between the two nucleons in (ds/2)2
is =152° while for (d^d^) it is =156°. Another way of saying this is that in the
/ = 1 states, the main difference between (d5,2)2 and (ds/2d3;2) is a flip of one
intrinsic spin—the matrix element corresponds to the strong35 interaction.
Therefore, we conclude that T = 1 configurations of identical nucleons are
 Collective Excitations in Even-Even Nuclei 145

not very mixed by short-range attractive residual interactions, because of the

large energy differences between such configurations, because of the conse-
quences of antisymmetrization in determining which spin states are allowed,
and because of the magnitude of the matrix elements that do exist. In contrast,
configurations of nonidentical nucleons can be strongly admixed. Moreover,
the mixing is a single nucleon effect. Therefore, such excitations cannot be
absorbed into an effective two-body interaction. The strong admixtures of dif-
ferent single nucleon wave functions, in this case ds/2 and d^, implies that the
spherical symmetry of the wave functions is lost since the resultant wave func-
tions must have nonuniform m state distributions (e.g.,(d5/2)2(p-n) has a com-
ponent M = 5/2 + 5/2 = 5 while (dM dM)(p-n) has Mmai = 5/2 + 3/2 = 4. Thus,
one can write the single-particle nuclear functions as i^= C5/2mds/2 + Cy^d^. As
we shall see, this is exactly the form of Nilsson wave functions for a deformed
shell model potential.
Although we have illustrated the idea for a particular case, the argument is
general. It is also interesting to note that the C.m coefficients must depend on
the substate m. In this example, Cm m=3/2 can be nonzero, but Cm m=5/2 must be
zero. Looking ahead for a moment, this, in essence, explains why Nilsson wave
functions are m-dependent (m is often called K in the Nilsson model) and also
why they are purest for the highest lvalues since, in that case, no other orbits
can contribute admixed amplitudes.
In closing these introductory pages it is worthwhile to re-emphasize that our
arguments for the existence of deformation and configuration mixing arose as
a consequence of the Pauli principle, which led to a different behavior of T=0
and T = I configurations, and of the short-range attractive nature of the
nucleon residual interaction. Nowhere was it necessary to specify the interac-
tion in detail. Of course, the choice of a specific residual interaction will affect
the detailed wave functions that result, but the possibility of nonspherical
wave functions is a rather general feature resulting from the particular con-
figurations and interactions allowed by the Pauli principle when nonidentical
nucleons are involved.

6.2 Collective Excitations in Spherical Even-Even Nuclei

One of the most characteristic empirical facts of nuclear systematics is that the
shell model picture of nearly independent particle motion under the influence
of weak residual interactions in simple configurations breaks down as one adds
more and more valence nucleons past magic numbers. Simply put, the residual
interactions among a growing number of valence nucleons build up to such an
extent that they obliterate much of the underlying shell structure. The shell
model wave functions become a poor first-order approximation to the real
nuclear wave functions. In short, they no longer serve as the most appropriate
basis states. In general, in a physical system, one always searches for some
suitable set of basis states such that the realistic wave functions are dominated
by one or a few components and any admixtures of basis states can be treated
as relatively small perturbations. This is not to say that the shell model cannot
provide a valid microscopic description of such collective excitations. Indeed,
146 Collectivity, Phase Transitions, Deformation

we shall see in a later chapter that the widely used and extremely important
RPA and TDA techniques are just such descriptions. Nevertheless, an alter-
nate viewpoint, that approaches the nuclear structure more macroscopically,
emphasizes the nuclear shape and excitations of that shape, providing a much
simpler, physically transparent approach.
In this chapter we shall discuss a sampling of the most important models for
collective excitations in even-even nuclei. As always, the emphasis will be on
the physical ideas.
To begin, we recall some of the systematics shown in Chapter 2. Figure 2.8
showed the energy levels of the Sn, Xe, Te, and Cd nuclei. Sn, with Z = 50, is
singly magic and displays a typical shell model behavior regardless of the
number of valence neutrons. The 2^ energy remains high and the 4^, 6^ levels
cluster. As soon as valence nucleons are added, for example in Te and Cd
(where the two valence protons are counted as holes), ZJ2j- drops sharply. The
decrease grows as the number of valence neutrons increases. The drop is even
faster for Xe, which has four valence protons. Figure 2.15 showed the sys-
tematics of the energy ratio E* {I E^\. It ranges from values < 2 for shell model
nuclei through ~2 for nuclei reasonably close to closed shells, then increases
sharply towards the limiting value of 3.33 near midshell. As we shall discuss,
values near 2.0,2.5, and 3.33 are all typical of different types of macroscopic
collective shapes: spherical harmonic vibrator, axially asymmetric rotor, and
axially symmetric rotor, respectively.
Generally, there is a smooth progression from one to another of these
idealized collective limits. However, inspection of Figs. 2.13 and 2.14 shows
that the systematics is anything but simple. At the end of this chapter we shall
see some easy, physically transparent, ways of understanding this complexity
and of parameterizing the behavior of heavy nuclei. Appropriately enough,
this approach will be based on a recognition of the importance of the residual
p-n interaction among the valence nucleons.
Here, though, we discuss models for each type of behavior, turning later to
their evolution from one into another. We start the discussion with the least
collective nuclei, which occur soon after closed shells: spherical-vibrational
nuclei. The generic concept of vibrational motion in nuclei is widespread and
encompasses a great richness of phenomena. We speak here of a particular
kind. To put this in context, suppose we expand the residual interaction among
the valence nucleons in multipoles, the first few terms will correspond to
monopole, dipole, quadrupole, octupole, and hexadecapole components. Each
of these carries a parity n= (-1)A where A is the multipolarity involved.
The electric dipole mode corresponds, geometrically, to a shift in the center
of mass, and therefore plays little role in the low-lying spectrum of even-even
nuclei. At higher energies, however, it induces the well-known giant dipole
resonance, which can be pictured as an oscillation of the proton distribution
against the neutron distribution. As this mode involves a rather large scale
displacement of major components of the nucleus, it requires considerable
energy, typically between 8 and 20 MeV. Since it is also a negative parity
excitation, and since most of the orbits in a given major shell consist of the
same parity, it necessarily involves excitations of the particles from one major
 Collective Excitations in Even-Even Nuclei 147

Fig. 6.1. The Ml scissors mode.

shell to the next and, once again, we see why it is generally high lying. We shall
not discuss the giant dipole resonance, or indeed, other giant resonances, any
further. This is in no way meant to minimize their importance: indeed, they are
a major focus of current work in nuclear structure physics. Their neglect here
stems rather from the emphasis in this book on the low-lying nuclear structure
spectrum and from the author's feeling that he has nothing particularly new or
innovative to say about the subject.
There is, however, a low-lying magnetic dipole excitation that has recently
been discovered in beautiful electron scattering and yray inelastic scattering
experiments. It occurs, for example, in heavy deformed nuclei at roughly 3
MeV and corresponds to a vibration in which the proton and neutron distribu-
tions osillate with respect to each other with a scissors type of motion, as
opposed to the linear vibrational motion of the giant electric dipole resonance.
The idea is illustrated in Fig. 6.1. This mode, characterized by strong Ml
electromagnetic transitions to the ground state and first 2* state is now known
in a number of nuclei and an interesting systematics has been established. It
has been studied from both geometric and algebraic (IBM-2) viewpoints.
Further discussion of this active area of research is beyond the scope of this
book.

Quadrupole Vibrations
The next vibrational mode, that we shall consider in detail, is the electric
quadrupole or E2 vibrational mode. It appears in different guises in different
categories of nuclei. Near closed shells, where the nuclei are spherical in their
ground state, the action of a quadrupole residual interaction causes the nu-
cleus to oscillate in shape, taking on a range of quadrupole distortions as a
function of time. The Hamiltonian for such a state can be schematically
written as

where Eg is the zero-point energy and the operators b1^ and b2^ create and
destroy this quadrupole vibration: y/^ = b^lO).
For simplicity of notation, and to keep the essential physics to the fore, we
148 Collectivity, Phase Transitions, Deformation

shall henceforth usually drop the subscripts on the operators b-. In the same
spirit, summations over the components /a will usually be implied rather than
explicit. Since we shall make frequent use of phonon or boson creation and
destruction operators, we pause for a moment to recall some key properties of
such operators in the formalism of second quantization. The basic defining
rules for arbitrary creation and destruction operators b and b f are:

and

where nb) is a state with nb bosons. Here b refers to quadrupole phonons;

later, in the discussion of the IBA, b will refer to either s or d bosons. From
these definitions

or

So, b% simply counts the number of b-type bosons. Thus we now recognize
that the second term in //is just the energy, relative to the ground state energy
E0 needed to create the quadrupole phonon excitation, which naturally carries
a spin and parity 2+.
There is no reason, except the limitations provided by the Pauli principle
when the microscopic structure of these vibrations is considered, that prevents
more than one phonon excitation from simultaneously existing. These mul-
tiphonon states y/N h = (b^lO) will correspond to higher and higher nuclear
levels. From Eq. 6.4, the second term in H is the product of the number of
quadrupole phonons and the energy of each. Clearly, at this stage in the
Hamiltonian one has a purely harmonic vibrational spectrum, where the
excitation energy is linear in the number of phonons: for an N A-phonon state,
Ex = Ha (Nph + 5/2), since the quadrupole mode is a 5-dimensional oscillator.

Table 6.1. m scheme for two-quadrupole phonon states. *

y1= 2 72 = 2
m /
i m2 M
-
2 2 4
2 1 3
2 0 2 4
2 -1 1
2 -2 0
1 1 2
1 0 1 2
1 -1 0
0 0 0 D 0
•Only positive total M values are shown: the table is symmetric Tor M < O.The full set of allowable m. values
giving M > 0 is obtained by the conditions ml > 0, m; < mr
 Collective Excitations in Even-Even Nuclei 149

Table 6.2. m scheme for three-quadrupole phonon states*

y, = 2 Jt = 2 73 = 2 M /
m
i m
i m,

2 2 2 6
2 2 1 5
"1
Z,

2 2 -1 3
2 2 -2 2
2 1 1 4
2 1 0 3
2 1 -1 2
2 1 _2 1
2 0 0 2
-1 •\ 1 1 A 1, £D
/. —1 1
—— n
2 0 -2 0
2 -1 -1 0 "
1 1 1 3
1 1 0 2
1 1 —1
1
1 1 —2 0
1 0 0 1
1 0 -1 0
0 0 0 0 13 o
7 = 6,4,3,2,0
"Only positive total M values are shown; the table is symmetric for M < 0. The full set of allowable m. values
giving M > 0 is obtained by the conditions ml a 0, m} S m2 < m t .

To continue, we must now turn to the question of which spin states are
allowed in multiphonon excitations. For the two-quadrupole phonon case, it is
clear that the maximum possible spin is 4+. But it turns out that only a triplet
of levels with spins J* = 0+, 2+, 4+ is allowed. There are many ways to derive this
result. Perhaps the most elegant is the use of Young tableaux, but here we shall
use the simpler and more straightforward, though more tedious, method of the
m-scheme. The essential difference between the use of the m-scheme for
phonon excitations and for single-particle excitations is the recognition that
phonons, involving particle-hole excitations and integer spins, behave essen-
tially as bosons. Therefore, the Pauli principle is not applicable and the wave
functions must be totally symmetric. This means that all combinations of m
states are allowed. Table 6.1 shows the m-scheme counting of substates for the
case of two quadrupole phonons and shows that the allowed spins are as stated
previously. The m-scheme analysis for the three-phonon case is given in Table
6.2, which shows that this excitation comprises a quintuplet of levels (at three
times the single phonon energy), with spins J* = 0+, 2+, 3+, 4+, 6+. This harmonic
picture of single- and multi-phonon excitations is illustrated in Fig. 6.2.
To pursue the study of multiphonon states, it is necessary to delve more
deeply into their structure. Consider the three-phonon levels. As Fig. 6.3
illustrates, the 6+ state can only be made in one way: by aligning the angular
momentum of a single phonon state with that of the 4+ two-phonon level.
150 Collectivity, Phase Transitions, Deformation

Fig. 6.2. Low-lying levels of the harmonic vibrator phonon model.

Similarly, the 03+three-phonon level can only be constructed by antialigning a

single quadrupole phonon with the 2+ two-phonon state. However, the other
three-phonon levels can be constructed in more than one way. For example,
the 2+ level can be made by coupling a quadrupole phonon with the 2+ two-
phonon state or by antialigning a quadrupole phonon with the 4+ two-phonon
state. In similar fashion, the 3+ and 4+ three-phonon states can be made by
coupling the third quadrupole phonon to more than one of the two-phonon

Fig. 6.3. Two-phonon composition of three-phonon states.

 Collective Excitations in Even-Even Nuclei 151

Table 6.3. Relative coefficients of fractional parentage for three-phonon quadrupole vibrator
states*

V2*
^ °* 4+
+
6 V3
4+ vn/7 VlO/7
3+ VT5/7 -V6A7
2+ V7/5 V4/7 V36/35
0+ VI
The normalization is to Vjvf For the three-phonon states, / the squares of the coefficients for each / also
give the relative values of B(E2:/3 —^Ji2 ). For example, the B(E2:4*3. —»2 + 2 .) and B(E2:4+ — > 4 + ) values are
in the ratio 11/10 = 1.1. See text.

states. The wave functions for the 2+, 3+, and 4+ three-phonon states are
therefore linear combinations of two terms, and the relative amplitudes are
simply phonon coefficients of fractional parentage whose squares give the
relative likelihoods that the three-phonon state is made in a certain way. It is
often useful to know these coefficients, so we give them for the N^ - 3 states
in Table 6.3. We will encounter two applications of these coefficients momen-
tarily.
An important aspect of the vibrational model centers on electromagnetic
transition rates since they are particularly sensitive to coherence properties in
nuclear wave functions. We saw, for example, in Fig. 2.16, the systematics of
B(E2:Oj+ -> 2j+) values throughout the periodic chart. Although small and
comparable to single-particle estimates in light nuclei, they attain values
orders of magnitude larger in heavy deformed regions. Intermediate values
characterize the realm of spherical-vibrational nuclei we are presently consid-
ering.
In general, radiation can be given off when any nucleon changes its orbit.
For example, changes in single-particle orbits in shell model nuclei are often
accompanied by the emission of y-radiation. While collective excitations are
clearly not of single-particle nature and the destruction of one does not
correspond to a single change of orbit by an individual nucleon, we will see in
Chapter 9 that their wave functions can be represented as coherent linear
combinations of single-particle-hole (or, equivalently, two quasi-particle)
excitations. Therefore, not only are y-ray transitions between phonon levels
permitted, but the coherence can make them particularly strong. Since a two-
phonon excitation involves a superposition of two linear combinations of one-
body excitations, the destruction of two-phonons would require a simultane-
ous destruction of two particle-hole excitations or four quasi-particles. There-
fore, such transitions are forbidden and one has the characteristic phonon
model selection rule AW h = ±1, where N h is the number of phonons. The
argument for this selection rule (obtained here for quadrupole vibrations of
spherical nuclei) is rather general and applies to any phonon structure de-
scribed as a linear combination of one-particle excitations. We will encounter
it repeatedly in various applications.
152 Collectivity, Phase Transitions, Deformation

Fig. 6.4. B(E2) values in the harmonic vibrator model.

Let us now consider the magnitude of these B(E2) values between phonon
states as illustrated in Fig. 6.2 and 6.4, where we assign a value of unity for the
decay of the one-phonon 2+ state to the ground state. Since, in first order,
multiphonon excitations simply consist of the piling on or superposition
of more than one identical phonon, it might seem that the B (E2) value
for the decay of the two-phonon state would also be unity. However, this
neglects the fact that there are two phonons in the initial state and that either
one of them may be destroyed. This gives twice as many decay possibilities and
therefore B(E2:(W = 2) -> 1) = 2, as indicated in Fig. 6.2. Continuing this, one
can state a general expression for the decay of the N ^-phonon state to the
(A^A-l)-phonon state. A transition N^ -> Wph-l must be accomplished by an
E2 operator of the form b, that is, a one-phonon destruction operator. By Eq.
6.2

and so the B(E2) value is proportional to AT .

This general statement, however, obscures the important point that, for
Nfh> 3, angular momentum conservation allows the decay of some initial
states to more than one final state. For example, the 2+, 3+, and 4+, three-
 Collective Excitations in Even-Even Nuclei 153

phonon levels can each decay to two or more of the two-phonon states. We
shall now show that the proportionality of these B (E2) values to the number of
phonons in the initial state actually refers to the sum of the B(E2) values from
a given Nflt-phonon level to all possible (N^-lj-phonon levels. It is trivial to
work out the relative B (E2) values for each of these decay routes: this exercise
is in fact one of the promised applications of the phonon CFP's in Table 6.3.
Consider as an example the decay of the N^ = 3,4+ level to the 2+ and 4+ two-
phonon states.
From Table 6.3, the wave function for the three-phonon 4+ level can be
written, in obvious notation, as

Then, the E2 matrix element connecting this level to the 2+ two-phonon state is

But the second term vanishes because and are orthogonal.

Hence, using Eq. 6.2 and setting E we get

Similarly,

and we see that the three-phonon —> two-phonon B(E2) values are propor-
tional to the squares of the three-phonon CFP coefficients in Table 6.3. The
table can be used to obtain the B(E2) values we have not worked out. The
results for the decay of the 4+ and 2+ three-phonon states are illustrated in Fig.
6.4. We also see an example of our general result, namely,

and similarly for the other three-phonon levels.

The reader is cautioned to bear these results in mind, since one occasionally
encounters statements such as that the B(E2) for the decay of a three-phonon
state to a two-phonon state is three times that for the decay of the one-phonon
state to the ground state. The proper relation involves the sum of the decays to
the possible final states.
This discussion of energies and B(E2) values in the harmonic vibrator
model assumes an idealized picture in which all of the phonons in a mul-
tiphonon state are identical. However, as we shall see in more detail when we
consider the microscopic structure of collective vibrations, a phonon state can
be written as a linear combination of individual particle-hole (or two quasi-
particle) excitations. Since a multiphonon wave function can be written as a
product of single phonon wave functions, multiphonon states effectively corre-
154 Collectivity, Phase Transitions, Deformation

spond to multiparticle-multihole excitations. Since each of these excitations is

a Fermion excitation, the Pauli principle must be obeyed. Its effects are
significant only when the particle-hole excitations cause significant occupa-
tion of individual / shells. As a trivial example, consider a nucleus where a
valence s1/2 orbit has, on average, one valence nucleon in the ground stale (i.e.,
the amplitude for (s1/2)2 is 0.5). Further suppose that a particular vibration has
an amplitude of 0.707 for raising a particle to this s1/2 orbit. Clearly, it is not
difficult to create a single- or even a double-phonon excitation in this case.
The latter would add, on average, one particle to the s1/2 orbit, thereby filling it.
This component of the vibrational wave function would be "blocked," how-
ever, in a three-phonon state, since one cannot put more than two particles in
an s1/2 orbit. The three-phonon wave function would therefore entail a modi-
fication to the basic phonon structure. This, in turn, would entail a change in
the B(E2) values for the decay of this state compared to the harmonic phonon
picture. Therefore, in realistic situations, the simple selection rules and the
analytic results for the relative B(E2) values between various phonon states
will be at best approximately realized.
Moreover, this blocking leads to anharmonic vibrational spectra in which
the degeneracies of the states in multiphonon multiplets are broken. Such
degeneracy breaking can also arise from the neglect of those residual interac-
tions not taken into account in the microscopic structure of the phonon itself.
In general, the calculation of such anharmonicities is complex and depends on
the specific/shells and residual interactions involved. However, there is one
situation in which a very simple and elegant result can be obtained essentially
by inspection.
This brings up the second application of the CFP's in Table 6.3. Suppose we
assume that degeneracy breaking is caused by residual two-body interactions
only. This means that the level energies in a two-phonon state are simply nol
twice the one-phonon energy, but differ because of a residual interaction
between the two-phonons. This, in fact, is the effect of the third term of H in
Eq. 6.1: it represents an interaction between two phonons.
In the three-phonon states, the same residual interactions apply and our
assumption simply states that there are no mutual interactions among the
three phonons at a given time. In this case, without ever specifying the residual
interaction, the nature, or the microscopic structure of the phonon, one can
immediately deduce the anharmonic energies of the three-phonon states from
those of the two-phonon levels. The situation is illustrated in Fig. 6.5, where
the energies of the 0+, 2+, and 4+ two-phonon states are written in terms of the
harmonic value 2£ 2 {, plus a perturbation e, (e0, e2, e4 are the anharmonicities).
Consider now the three-phonon levels. As we have seen, there is only one way
to make the 6+ level: by aligning a single 2+ quadrupole phonon with a pair of
phonons coupled so as to produce the 4+ two-phonon state. In the three-
phonon 6+ state, there are three possible pairs of (indistinguishable) phonons
that can couple and interact in forming the intermediate 4+ plus state. There-
fore, the anharmonicity (the deviation of £6+ from 3E2{ ) is three times the
anharmonicity in the 4+ two-phonon state, or 3e4. In the same fashion, the 0' 3-
 Collective Excitations in Even-Even Nuclei 155

Fig. 6.5. Energy anharmonicities in the vibrator model assuming arbitrary two-body residual
interactions.

phonon state can only be made by antialigning one phonon with the 2+ two-
phonon state. Again, there are three ways to do this, and the anharmonicity in
the three-phonon 0+ energy will be triple the anharmonicity in the two-phonon
2+ level, or 3e2. The other three-phonon states, which can be made in two or
more ways from the 2-phonon levels, will have total energy anharmonicilies
given by the relative proportions of their wave functions arising from the
various two-phonon states. These relative proportions are given by the CFP
coefficients of Table 6.3, and the resulting energy anharmonicities are shown
in Table 6.4 and illustrated in Fig. 6.5.
It is worthwhile to reiterate what has been derived here. We have never
specified the structure of the phonon itself. We have also never specified the
nature of the residual interaction except to state that it is two-body. Neverthc-
156 Collectivity, Phase Transitions, Deformation

Table 6.4. Energies of the three-phonon quintuplet states in terms of the two-phonon anhar-
monicities.
N. J Energy*
(relative units)
0 0 0
1 2 1
2 0 2 + efl
2 2 2 + e2
2 4 2 + E4
3 0 3+3e 2
3 2 3 + 7/5e0 + 4/7e, + 36/35e4
3 3 3 + 15/7£2 + 6/7e4
3 4 3 + ll/7£2+10/7e4
3 6 3+3£ 0

*£a, £[, and £4 are defined as the deviations of the 0*, 2*, and 4* level energies of the two-phonon triplet from

less, from the observed anharmonicities in the two-phonon states, we have

been able to derive predicted anharmonicities for the three-phonon levels.
This model can be tested by observation of the three-phonon levels. If
discrepancies are found, then we immediately know that they must arise either
from three-body interactions, from Pauli principle effects in the multiphonon
states, or from interactions with other nearby excitations. Our analysis helps
to isolate the specific mechanisms leading to the observed anharmonicities.
The vibrational model was originally proposed in the early 1950s, and it was
thought that numerous examples of such a structure were observed empiri-
cally. Triplets of levels with spins 0*, 2+, 4+, were known in many nuclei near
closed shells at slightly more than twice the energy (typically, E2ph/E^ph ~ 2.2) of
the first 2+ state. However, it was commonly found that the energy splitting
among the two-phonon levels is comparable to the one-phonon energy; basi-
cally, the structural effects leading to anharmonicities are comparable to those
involving the phonon itself. Moreover, the predicted phonon model B(E2)
values were significantly violated. Perhaps most importantly, over the last
decades, additional low-lying levels near the two-phonon states have been
detected in many nuclei.
While in many of these nuclei there may be an underlying vibrational
structure, it is significantly perturbed and admixed with other degrees of
freedom. An interesting example is the Cd isotopes shown in Fig 6.6: These
nuclei, especially 114Cd, historically have been considered the best prototypes
of vibrational behavior. However, it is clear from Fig. 6.6 that there are four
and sometimes five levels clustered together in the two-phonon energy region.
The extra states have since been identified as intruder levels.
It is worth digressing from our discussion of vibrational excitations to
comment on such intruder excitations, which are now known to be widespread
in nuclei near closed shells. They are an area of active study, especially in
 Collective Excitations in Even-Even Nuclei 157

106 IO8 110 112 111 116 118 120 122 124

CADMIUM SYSTEMATICS

Fig. 6.6. Energy systematics in the Cd isotopes. Note the intermingling of two-phonon triplet
states with extra levels (Aprahamian, 1984).

nuclear regions far from stability. An understanding of them is, in any case,
essential to disentangling the empirical features of many vibrational nuclei.
Moreover, the discussion will introduce some basic ideas relating to the p-n
interaction, which we shall return to later in this chapter. We have already
alluded several times to the idea that this interaction is essential to the
development of collectivity and deformation in nuclei. Here we will encounter
our first specific example of this.
Cd has Z = 48, and therefore two proton holes relative to the Z = 50 magic
number. It is of course possible to excite the Cd nuclei by elevating two
protons from the Z = 28-50 valence shell into the next higher, Z = 50-82 shell.
The idea is sketched in Fig. 6.7. Normally this requires considerable energy
since it involves raising two nucleons across a major shell gap. However, the
residual p-n interaction is strong and attractive, and as such introduces a major
modification to this first-order energy. In a simple picture, one can view the
"normal" states of Cd as consisting of two proton holes interacting with some
number of valence neutrons. In the intruder state, there are four proton holes
in the Z = 28-50 shell plus two proton particles in the Z = 50-82 shell. In a
sense, there are six valence protons that can now interact with the same
number of valence neutrons. In this picture the intruder states in Cd are
analogous to the normal states in Ba, as suggested in Fig. 6.7.
We have already seen that the more valence nucleons there are of both
kinds, the "softer" the structure will be. Sufficient numbers of valence protons
and neutrons lead to deformed shapes. Therefore, in this schematic view,
intruder levels should be more deformed than the "normal" levels. Moreover,
since the attractive interaction is three times greater than in the normal levels,
158 Collectivity, Phase Transitions, Deformation

INTRUDER STATE MODEL

PROTON LEVELS

Fig. 6.7. Schematic illustration of intruder excitations and normal states in Cd. The normal
states of Ba are shown for comparison with the Cd intruders.

the intruder state excitation energies are lowered relative to their unperturbed
(no p-n interaction) value. This lowering increases approximately linearly
with the number of neutrons, and therefore, one expects the energies of these
intruder states to drop from approximately twice the shell gap for 7V-50 or 82
toward midshell, where the p-n interaction strength is maximum. Inspection
of Fig. 6.6 shows that this simple picture is at least qualitatively correct.
As suggested, this intruder state model is quite general and such excitations,
once thought to be rare, are now known to abound throughout the periodic
table. Perhaps the best known example is in the Pb region, whose systemalics
are shown in Fig. 6.8.
Most intruder levels observed to date are proton excitations. This is related
to the role of a strong p-n interaction in lowering these levels. Since there is a
 Collective Excitations in Even-Even Nuclei 159

Fig. 6.8. Systematics of 0+ intruder levels in the Pb isotopes (Van Duppen, 1985).

neutron excess in heavy nuclei, the excited valence protons in the intruder
state occupy the same shell as the neutrons, thus enhancing the p-n interaction
(see Fig. 3.5).
The concept of intruder states is far more important than the explanation of
a fewbothersome levels. It is closely connected with the origin of deformation
itself, as we shall discuss towards the end of this chapter.
These comments relate to our discussion of Cd vibrational states because,
as the TV = 82 shell closure in Cd is approached, the expected rise in intruder
energies should leave behind a reasonable vibrational spectrum. This has led
to experiments on 118>120Cd. The data included in Fig 6.6 seem to confirm this
expectation, although studies of absolute B(E2) values show that the interpre-
tation is not quite so simple. The level scheme for 118C is illustrated in Fig 6.9,
which includes the known information on y-ray transition rates and relative
B(E2) values. There is a triplet of levels near 1200 keV in which the energy
separation is much less than Ei\. Furthermore, and most remarkable, an
entire closely spaced quintuplet of candidates for the three-phonon multiplet
was identified. (In fact, candidates for even higher four-phonon states have
been suggested.) Although there are significant deviations from the expected
160 Collectivity, Phase Transitions, Deformation

Fig. 6.9. Level scheme of 118Cd showing the one-, two-, and three-phonon states as well as an
intruder (T level at 1615 keV and possible candidates for four-phonon excitations above 2.2 MeV.
On the right are shown the average (A/V h = 1)/(A/V h = 2) branching ratios. On the left are the
predictions for the three-phonon states assuming the empirically observed anharmonicities in the
two-phonon states. These are the same predictions one would obtain in the U(5) limit of the IB A
(Aprahamian, 1987).

patterns of relative B(E2) values compared to the phonon model predictions,

the overall predominance o(ANfti = 1 transitions is well satisfied. On the right
of Fig 6.9, the average ratio of one-phonon to two-phonon changing transitions
[(AAf A = l)/(AAf A = 2)] is indicated: there is at least an order of magnitude
preference for the phonon model selection rule. On the left are the predicted
energies for the three-phonon quintuplet based on the anharmonicities ob-
served in the two-phonon triplet. Although the agreement with experiment is
not at all exact, the observed clustering of the levels into two spin groups is
 Collective Excitations in Even-Even Nuclei 161

correctly predicted. Presumably, the remaining deviations are due in part to

intruder-normal state mixing or to Pauli principle effects. In fact, calculations
by Heyde and co-workers that incorporate such mixing provide a significant
improvement in the predicted energy levels.
The observation of E2 transitions in Cd, as well as other near-vibrational
nuclei, such as the weak ANf/t - 2 transitions that violate the phonon model
selection rules are often considered to arise from interactions that mix the
number of phonons. From the B(E2) branching ratios, one can make esti-
mates of the mixing amplitudes and thereby extract the interaction matrix
elements. A similar approach can be used for intruder states. These states
have forbidden E2 transition matrix elements to the normal states, since they
cannot be connected to them by a one-particle operator. Once again, however,
neglected residual interactions will cause some small mixing of the intruder
and normal states, and such mixing can be probed empirically by measuring E2
branching ratios. It is generally estimated that the intruder-normal mixing
matrix elements are -100 keV. Therefore, in a nucleus such as 1I4Cd where the
final energies of the intruder and normal 0+ states near 1.2 MeV are only -200
keV apart, these states must have been nearly degenerate prior to mixing
(after mixing, degenerate levels are twice the mixing matrix element apart),
and therefore their perturbed wave functions will be nearly equal admixtures
of the unperturbed intruder and normal states. In 118Cd, on the other hand, the
final separation, and therefore the initial separation, is much greater; the
mixing is correspondingly smaller.
There are other nuclei that display reasonably well-developed vibrational
spectra. Examples are 102Ru, which has a full three-phonon quintuplet but with
greater intramultiplet splitting (anharmonicity) than 118Cd. The Te isotopes
near A = 120 also exhibit nice two-phonon triplets, relatively closely bunched.
The level schemes for two of these nuclei were included in Fig. 2.9.
Despite these examples, near-harmonic vibrational motion is the exception
rather than the rule. The reason seems to be that it takes only a few valence
protons and neutrons to soften the nucleus to deformation to such an extent
that the simple scheme of quadrupole surface vibrations of a spherical shape
loses applicability.

6.3 Deformed Nuclei: Shapes

Further from closed shells, the accumulating p-n interaction strength leads to
additional configuration mixing and deviations from spherical symmetry even
in the ground state, and so we now turn to consider nuclei with stable and
permanent deformations. The lowest applicable shape component is a quad-
rupole distortion. There can also be octupole and hexadecapole shapes.
These are schematically illustrated in Fig. 6.10a. Nuclei with quadrupole
shapes abound throughout the periodic table in midshell regions.
For a nucleus with quadrupole deformation, one can write the nuclear
radius as
162 Collectivity, Phase Transitions, Deformation

Fig. 6.10. (a) Equal potential surfaces for different multipole distortions, (b) Schematic illustra-
tions of various quadrupole shapes (prolate, oblate, axially asymmetric) as well as of /and /3
vibrational motions.

where R0 is the radius of the spherical nucleus of the same volume.

The Y2 are spherical harmonics of order 2 and the a are expansion
coefficients. It is convenient to change notations here and write the five a in
terms of three Euler angles and two variables p and y. We set a^ = a^ = Q, since
 Collective Excitations in Even-Even Nuclei 163

these two coefficients represent only the motion of the nuclear center of mass,
and write a0 = /fcosyand c^ - a2 = /feiny. The nuclear shape is then specified
in terms of ft andy. j3 represents the extent of quadrupole deformation, while
ygives the degree of axial asymmetry. Most nuclei are axially symmetric, or
close to it, at least in their ground states. For an axially symmetric nucleus, the
potential has a minimum at y= 0°. [It is unfortunate that no single notation for
deformation parameters exists. /? is quite common, but we shall also encounter
e and 8, especially in Chapters 7 and 8. Often, a subscript "2" is appended to
explicitly denote quadrupole deformation.]
The relation between fi, y, and the nuclear radii can be seen by evaluating
the change in radius (Rx z- R0) in Cartesian coordinates as a function of j3 and

To see the shapes implied by these expressions, Table 6.5 gives the values of
these correction terms to a spherical shape for four y values in units
of f5/47r RJ5- Values greater than zero in the table indicate an elongation in
the direction concerned; those less than zero indicate a compression. Note
that, for y values that are a multiple of 60°, two Rvalues are always identical
since the nucleus is axially symmetric for these y values. For y= 0°, the nucleus
is extended in the z-direction and compressed in x and y. This is a prolate
(American football) shape. Oblate (disk-like) nuclei correspond to y= 60° and
180° and are compressed in the y- and z-directions, respectively, and extended
in the xz and xy planes, respectively.
The essential difference between prolate and oblate shapes is that the form-
er is extended in one direction and squeezed in two, while oblate shapes are
extended in two and compressed in one. Intermediate values of y(y* nnlY),
such as the y = 30° example in the table, correspond to axially asymmetric
shapes, that is, to a flattening of the nucleus in one of the two directions
perpendicular to the symmetry axis. Then all three radii are different.
An attempt has been made to depict several nuclear shapes in Fig. 6.10b (as
well as [} and y vibrational motions to be discussed). These pictorial images,

Table 6.5. Changes in the radius of a quadrupole ellipsoid in the x, y,, z directions for several y
values and fixed /}. *

r
0° 30° 60° 180°
SK
< +1 +0.866 +1/2 -1

«;
8R -1/2 0 +1/2 +1/2
-1/2 -0.866 _] +1/2
*AU numbers are in units of VS/4^R0/3.
164 Collectivity, Phase Transitions, Deformation

Fig. 6.11. Empirical systematics of quadrupole deformation parameters /3 in the rare earth region.

while crude and too classical, should be helpful to readers unfamiliar with the
shapes involved.
The systematics of Rvalues (effectively, quadrupole moments) for the rare
earth region is shown in Fig. 6.11. The qualitative behavior is easily under-
stood in terms of a generalization of the seniority argument of Chapter 5 (see
Eq. 5.17). Early in a major shell, when softness to deformation first appears,
the individual; orbits are still nearly empty; hence the quadrupole moments
for the nucleons in these orbits are positive. Then a large positive Q(ft) builds
up rapidly. As the shell fills, however, the contribution of successive /' shells to
the total quadrupole moment decreases, vanishes, and ultimately turns nega-
tive (see Fig. 5.4). On account of these negative contributions, the summation
over the individual quadrupole moments steadily decreases and may even go
negative (as in Pt, not shown in Fig. 6.11) near the end of the shell.
Two important quantities for a quadrupole deformed nucleus are the
moment of inertia and the quadrupole moment of the ellipsoidal shape. Both
can be written in terms of ft for axially symmetric nuclei. For an ellipsoid, the
so-called rigid body moment of inertia is / = 2/5 Mi2. Integrating the radius
over the nuclear surface gives (to first order in ft)

The intrinsic quadrupole moment is given by

to second order in p.
Note that since Ra <* /I1'3, / = /I5'3 and is a linear function of ft. Qa is
directly proportional to ft in leading order. For the ft values typical of actual
deformed nuclei, ft- 0.3, the higher-order terms are rather small.
 Collective Excitations in Even-Even Nuclei 165

6.4 Rotations and Vibrations of Axially Symmetric Deformed Nuclei

The most obvious characteristic of nonspherical nuclei is that they can un-
dergo rotations about an axis perpendicular to the symmetry axis. They can, of
course, also vibrate and, moreover, rotations can be superimposed on vibra-
tional motion. In order to discuss both rotational and vibrational motion, and
later, their superposition and interactions, we need to specify the deformed
wave functions.
These wave functions incorporate the two aspects of intrinsic excitation and
rotational motion. The latter is specified in terms of the well-known rotational
D matrices, the former in terms of the wave functions %K. The adiabatic
assumption of the separability of rotational and intrinsic motions leads to a
product wave function in D and %. In a state vo m a spherical nucleus, all
magnetic substates are degenerate since there can be no distinction in energy
as a function of angular orientation of the motion. This, of course, is not true
for deformed nuclei whose energies depend on the orientation of the wave
function with respect to the symmetry axis. However, whether we consider the
motion of a single nucleon in a deformed field (as we shall do when we discuss
the Nilsson model in the next chapter) or the motion of some collective "wave"
around the nucleus, there remains a twofold degeneracy, corresponding to
clockwise and counterclockwise motions, which persists even in the deformed
field. These two motions can be distinguished by the projection, K, of their
angular momenta on the symmetry axis. States with projections K and -K will
still be degenerate. The nuclear wave function must reflect this and thus one
has the symmetrized product form for wave functions in rotational nuclei:

Note that for K - 0, only even J values are allowed, so wave function collapses
to a single term

With these wave functions in hand, we consider axially symmetric nuclei in

which the rotation has equal frequencies around the x or y axes. The rotational
Hamiltonian is simply

where / is the moment of inertia and R is the rotational angular momentum

operator. If we assume that the ground state is f = 0+, K = 0, and if all the
angular momentum can be ascribed to rotation (as is normally true for the low-
lying, low-spin, positive parity states in deformed even-even nuclei) then the
total angular momentum J = R and we obtain the famous symmetric top
rotational energy expression

where only even./ are allowed.

166 Collectivity, Phase Transitions, Deformation

It is useful to comment briefly on the assumption of a 0+ ground state. For

spherical nuclei, we have seen that a 5-interaction between two identical
particles in a j shell produces a 0+ ground state. For the case of multiparticle
configurations of identical nucleons in the same / shell, the seniority v = 0 0+
state also emerges as the lowest energy configuration. This behavior does not
result from the special character of the interaction, but is characteristic of any
short-range interaction. The same result applies when one has both valence
neutrons and protons. It also occurs if/ is not a good quantum number, as when
there is single-nucleon configuration mixing. (As we shall see, such mixing
characterizes Nilsson model wave functions, which can be written as coherent
sums over several/ values.) The reason is that a short-range interaction favors
the lowering of the 0+ states in each /-configuration. The presence of many /
values actually serves to further lower the 0+ state because of a build up of
coherence. We have already seen this coherence effect in our discussion of
pairing, in which the lowering of the 0+ state is related to the size of the energy
gap A, which in turn depends (see Eq. 5.22) on the number of partially filled
orbits near the ground state and the strength parameter G. Thus the phenome-
non of a 0+ ground state persists in deformed even-even nuclei as well.
The symmetric top expression for rotational energies gives the values
(where we abbreviate K-I2I by #):

and so on. Thus, the energy ratio E^lEz^ = 3.33. This simple formula is one
of the most famous results of the rotational model and still remains one of the
best signatures for rotational motion and deformation. We have already seen
examples of nuclei that behave according to this relation in Figs. 2.10 and 2.15.
Combining Eqs. 6.8 and 6.13 gives two characteristic features of transitional
and deformed nuclei. For a given mass region (A = const), h2/2I decreases as
j3 increases, leading to smaller and smaller rotational spacings as a deformed
region is entered. This behavior is one of the signatures of nuclear transition
regions, as we pointed out in Chapter 2 (Figs. 2.13 and 2.14). Second, since
nuclear radii increase as Av3, I = A5'3 and W1I = A~s'3, for constant ft.
Rotational spacings should therefore decrease for heavier nuclei. Figure 2.10
illustrated exactly this behavior with examples taken from Zr, the rare earth,
and the actinide regions.
As we stated earlier, in a geometrical picture (that is, a macroscopic one in
which we do not worry explicitly about the Pauli principle), there is no reason
why rotational motion cannot be superposed on intrinsic excitations, whether
of collective vibrational or two-quasi-particle character. Now consider such an
intrinsic excitation in a deformed nucleus.
Each intrinsic excitation carries intrinsic angular momentum Jo. It can be
partially characterized by the projection of that angular momentum onto the
symmetry (z) axis. Since for axially symmetric nuclei, any rotation of the
 Collective Excitations in Even-Even Nuclei 167

nucleus as a whole must be about an axis perpendicular to the z axis, such

rotation has vanishing projection along the z axis. Therefore, the projection of
the total angular momentum / along the z axis, denoted K, is the same as that
of the intrinsic excitation. One sometimes sees the notation Q, for the projec-
tion of the intrinsic angular momentum. However, the assumption of axial
symmetry is generally a good approximation and AT is often used interchangea-
bly for both the projection of the intrinsic and total angular momenta. We shall
follow this simplified notation here.
When nuclear rotational motion is superimposed on an intrinsic excitation
characterized by projection K, the total angular momentum can take on values
/ = K, K +1, K + 2,..., except when K = 0, in which case only even spins J are
allowed. Thus, for the rotational energies, relative to the "base" energy of the
intrinsic excitation, Eq. 6.13 becomes

The energy expressions, Eqs. 6.13 and 6.14, are quite accurate for low spin
states in deformed nuclei, thus affirming the basic validity of the rotational
concept. An example is shown in Fig. 6.12 for a typical deformed nucleus, 164Er
(the nature of the intrinsic excitations indicated in the figure will be discussed
momentarily). The energies for each band are normalized to the bandheads in
order to isolate the rotational behavior. The predictions from Eq. 6.14 are
reasonable, but there are also clear deviations as J increases. Also, note the
changes in the inertial parameter, h 2/27, from band to band. Apparently the
deformation is not completely constant.
An understanding of the physics involved in these deviations is extremely

Fig. 6.12. Ground, /, and J3 band levels of a typical deformed nucleus 164Er. For each band the
symmetric top rotational energy predictions (Eq. 6.14) are shown.
168 Collectivity, Phase Transitions, Deformation

simple, yet intimately connected with a number of subtle effects involving

rotation-vibration coupling, bandmixing, axial asymmetry, and ^softness.
These concepts, and the various models emphasizing different aspects of them
(e.g., the Davydov and Wilets-Jean pictures), are so interrelated that a logi-
cally ordered pedagogical treatment is difficult. We have chosen to first
discuss the basic "Bohr-Mottelson" idea of an axially-symmetric, deformed
nucleus susceptible to quadrupole vibrations of so-called /? and /type. This
will allow us to proceed to consider their interactions with the ground state
band and with each other (that is, bandmixing or rotation-vibration coupling).
This, in turn, will provide us with a refinement to Eqs. 6.13 and 6.14. With this
in hand, we will turn to approaches that provide closely related viewpoints. 1 n
the end, we hope that both the different starting points and the intimate
relationships of these models will be clear.
We have seen that the most common distortion of spherical nuclei is
quadrupole in nature. Therefore, it should not be surprising that the most
common low-lying vibrational excitations in deformed nuclei are quadrupole
vibrations. Clearly such modes, which carry two units of angular momentum,
can be of two types with K = 0 and K = 2. The former are known as ft vibrations;
since K = 0, the vibration is aligned along the symmetry axis and therefore
preserves axial symmetry.
The latter, with K = 2, is called a 7 vibration and represents a dynamic time-
dependent excursion from axial symmetry. These names stem from the fact
that the /^vibration corresponds to fluctuations in the quadrupole deformation
J3, while the 7 vibration corresponds to oscillations in 7. A qualitative depic-
tion of the p and 7 modes is included in Fig. 6.10. The 7 vibration is more
difficult to visualize. It may be viewed as an alternate "squashing" of an
American football in two directions 90° to the symmetry or major axis. Note
that, although the 7 vibration involves a dynamic fluctuation in 7 and has an
average value of 7ave = 0°, the rms value of 7 is finite and can be quite large.
These two vibrational modes appear in essentially all deformed nuclei, usually
around 1 MeV for A > 100. The fact that they occur well below the pairing gap,
2A ~ 1.5 - 2 MeV, is not a violation of the concept of pairing correlations but an
affirmation of the collective character of these excitations (as we shall see
explicitly in Chapter 9). An example of a typical deformed nucleus with /} and
7 vibrations was shown in Fig. 6.12. Figure 2.17 showed the systematics of the
lowest-lying intrinsic K - 0 and K = 2 excitations in deformed rare earth nuclei.
Although it is possible that such excitations may involve components from
configurations other than /? or 7 vibrations (such as low-lying "pairing" vibra-
tions or two-quasi-particle excitations), it is safe to assume that nearly all the
excitations included in these figures are predominantly of ft or 7 type.
After our discussion of the microscopic structure of /? and 7vibrations in
Chapter 9, we will understand the origin of these systematics. For now we
simply treat these vibrations as phenomenological macroscopic shape excita-
tions. As we have noted, each of them can have rotational motion superposed.
Therefore, we can look on a level scheme such as that in Fig. 6.13 as a
prototypical one for a heavy, axially symmetric (or nearly so) deformed nu-
 Collective Excitations in Even-Even Nuclei 169

Fig. 6.13. Positive parity levels of a typical deformed nucleus.

cleus, with ground, j8 and 7 excitations, and rotational bands. Two-quasi-

particle excitations, each with its own rotational band, can also appear above
the "pairing gap", as indicated schematically. (Negative parity (octupole)
vibrations can also occur low in energy and will be briefly mentioned later.)
It is also possible to have multiphonon deformed vibrations. Since K values
are projections of the angular momentum and not themselves vectors, the K
values for multiple phonon excitations are obtained by simple algebraic sums
and differences of the component K values. Thus, the double j3 vibration has
K = 0, the /3 7 vibration has K = 2, and the double 7 vibration exists in two
forms with K = 0 and K = 4, but not K = 2.
In contrast to spherical vibrational nuclei where one quite frequently
encounters at least some two-phonon and occasionally three-phonon levels,
albeit with anharmonic distortions, deformed multivibrational states (77, j3j3, or
fly) are the exception rather than the rule. (Extensive studies of a few
deformed nuclei, such as 168Er and some of the Dy isotopes, have identified
candidates for multiphonon vibrations, but they are probably admixed with
other configurations and the data are both fragmentary and ambiguous.) One
reason for the dearth of multiphonon vibrations in deformed nuclei is that,
while single-phonon spherical vibrations typically occur at about 500-600 ke V,
placing the two-phonon states around 1.2 MeV, j3 and 7 vibrations are typically
at ~1 MeV, which would put the two-phonon states above 2 MeV. But the
energy gap 2A is also on the order of 2 MeV, so that there is a plethora of two-
quasi-particle excitations at the same energies. This fragments the collective
states and makes them harder to detect.
170 Collectivity, Phase Transitions, Deformation

In addition to the experimental problem of identifying multiphonon vibra-

tions in deformed nuclei, there is the question of the effects of the Pauli
principle. This was discussed for quadrupole vibrations of spherical nuclei—
the same arguments apply here, except more so. The reason is that, in
spherical nuclei, a given single-particle level such as L,a can contain up to eight
particles, and the Pauli principle will not play a large role if the level is less than
half filled. In deformed nuclei, however, each intrinsic excitation (Nilsson
orbit with given K value) is only two-fold degenerate (see Chapter 7): if the
excitation is important in the one-phonon state, there will be a substantial
"blocking" effect due to the Pauli principle in a two-phonon vibration. De-
spite extensive recent discussion, the issue of the inhibition of multiphonon
excitations in deformed nuclei is not yet settled. Calculations testing the
possibility of such excitations involve large bases that incorporate both quasi-
particle and collective degrees of freedom and, in most cases, they are simpli-
fied by truncating the space. Different truncations schemes yield different
results.
Perhaps the most telling and interesting properties of the /? and 7 vibrations
centers on their electromagnetic decay properties. The basic E2 selection rule
here is identical to, and arises from the same arguments as in the phonon case.
Microscopically, the /3 and 7 vibrations can be written as linear combinations
of two-quasi-particle excitations (or, in nonpairing terminology, particle-hole
excitations). Therefore, an electromagnetic transition can create or destroy at
most one such vibration or phonon. A /3 or 7 vibration can decay, therefore,
by E2 radiation to the ground state band, but transitions between j and /3
vibrational bands are forbidden since they involve the simultaneous destruc-
tion of one vibration and creation of another. For multiphonon vibrations (yy,
ftp, /ty) this selection rule allows 77-* 7, /J/7 -> ft and /3y-» 7 or ft transitions and,
indeed, such transitions are among the key signatures used in searching for
such excitations.
We shall see in Chapter 9 that the "collectivity" of /? and 7 vibrations is such
that their wave functions typically involve a small handful of orbits comprising
a subset of the valence particles. Since collectivity in electromagnetic transi-
tions arises from coherence in the wave functions, we can expect that 7—> g or
/? -> g E2 transitions will be collective (the matrix elements will be much larger
than single-particle matrix elements), but that they will be much weaker than
transitions occurring within a given rotational band since a change in rota-
tional structure involves the whole nucleus (or at least all of the valence
nucleons).
We saw in Chapter 2 that rotational transitions in even-even deformed
nuclei, typified by B(E2: 2,+ -» Ot+) values, can easily reach several hundred
single-particle units. At the same time, vibrational transitions such as 7~> g are
typically 10-30 single-particle units. As we shall see in Chapter 9, while it is
relatively easy to construct K=2 two-quasi-particle states by breaking nucleon
pairs and exciting one particle to an excited quasi-particle level, it is not so easy
to create K = 0 excitations. This normally involves the excitation of a pair of
nucleons together. It should not be surprising that B(E2) values for /3 —> g
 Collective Excitations in Even-Even Nuclei 171

transitions, while collective, are much weaker than y-» g transitions, typically
a few single-particle units. We illustrated these points in Chapter 2 (Fig. 2.18)
by summarizing the transition rate data for deformed nuclei in terms of ratios
of y-> g and /? -> g B(E2) values to B(E2:2* -> Oj+) values. One remarkable
feature is the relative constancy of the y-» g B(E2) values. This surely points
to a collective, slowly evolving structure.
To further consider y-ray transitions, both within rotational bands (intra-
band transitions) and between intrinsic excitations (interband transitions), we
must be a bit more formal. The basic results are extremely simple to derive.
Taking the wave function of Eq. 6.10 for deformed nuclei, the E2 transition
matrix element (up to constant factors) is:

where we have neglected the cross terms of the form <JOE2II-.K.), which
normally vanish by conservation of angular momentum, and have utilized the
fact that the Kf -> K. and -Kf -» -K. matrix elements are identical. The
separation of the wave function into rotational and vibrational components
thus gives a separation of the matrix element into an intrinsic part (second
factor on the right in Eq. 6.15) dependent only on % and a "rotational" part,
which in turn depends only on the angular momenta involved and is propor-
tional to the Clebsch-Gordon coefficient (JK2 AK\JKf}.
The diagonal matrix elements with j. = %f, J. = /.give the intrinsic quad-
rupole moments Qo of the excitation involved. Note that these quadrupole
moments are given in the intrinsic body-fixed frame. The observed quad-
rupole moments, that is, the so-called spectroscopic quadrupole moments,
involve a transformation to the laboratory frame, giving the well-known result

The dependence on K and / reflects the fact that the perceived shape of a
rotating nucleus is not the same as the shape in the intrinsic frame. This is easy
to visualize. When a prolate deformed nucleus rotates about an axis perpen-
dicular to the symmetry axis, the time averaged shape looks more like a disk
(an oblate nucleus), which would have a quadrupole moment of the opposite
sign. This effect is exacerbated for higher rotational velocities and, indeed, for
J(J +1) > 3K2, the spectroscopic quadrupole moment does have a sign opposite
to the intrinsic quadrupole moment. In fact, for K = 0 this is always the case.
Note that for J = 0 (which implies K = 0 since K < J), Q = 0: a state of zero
angular momentum can have no preferred direction of the time averaged
distribution in space and therefore no quadrupole moment.
For the important case of matrix elements diagonal in % but not in J (or
transitions within a band), we have nearly the same result except for a Clebsch-
Gordon coefficient connecting 7. and 7- Thus
172 Collectivity, Phase Transitions, Deformation

Since the intrinsic quadrupole moment Qo<* P(l + 0.16/3), B(E2:0^ -^ 2/) ~ I?.
The large Rvalues (of about 0.3) that characterize deformed nuclei can lead to
a one to two order of magnitude increases in this B(E2) value above that of
near spherical nuclei (j3 ~ 0.05). This explains the systematics we showed in
Fig. 2.16, which provides the most obvious evidence of deformed collective
behavior in nuclei.
Frequently one can extract very sensitive and critical information on struc-
ture effects and rotation-vibration interactions from B(E2) values for a pair of
transitions connecting the same two intrinsic states. These transitions can be
either both intraband or both interband with the same initial and final bands.
Then the intrinsic matrix element will clearly be identical for both transitions
and will cancel in their ratio. Such branching ratios depend only on the squares
of Clebsch-Gordon coefficients, and are therefore model independent in the
sense that they do not depend on the microscopic structure of the excitations
involved. They depend on the assumption of the separability of rotational and
vibrational motions. They are known as Alaga rules.
Specifically, we have

Note that, since the intrinsic structure has canceled out in such ratios, they
are equally valid for transitions involving any intrinsic states (e.g., two-quasi-
particle states) as well as for those involving vibrational excitations.
As examples of these ratios, we have

Equally simple but numerically different results are obtained for other
transitions. There are two important uses of such ratios. First, since they
depend on the K values of the initial and final states, they can sometimes be
empirically used to assign K quantum numbers to different intrinsic excita-
tions. Secondly, as we shall see momentarily, small admixtures of different
intrinsic excitations (bandmixing effects) can induce enormous changes in
these branching ratios, so the empirical ratios can provide very sensitive tests
of small details of the nuclear wave functions.
It is easy to look up or calculate values of the Clebsch-Gordon coefficients
 Collective Excitations in Even-Even Nuclei '• 173

Table 6.6. Some useful Alaga rules for E2 transitions in deformed nuclei*
(J.K2AK \fffi
/. J
f K^Kf
0->0 2->0 0-»2 2->2
0 2 1.0 1.0 —
—
2 0 0.200 0.200 — —
2 0.286 0.286 0.286 0.286
3 — — 0.500 0.500
4 0.515 0.014 0.215 0.215
3 2 — 0.358
— 0.358
3 — — — 0
4 — 0.143 — 0.343
5 — — — 0.300
4 2 0.286 0.120 0.008 0.120
3 — — 0.112 0.267
4 0.260 0.351 0.351 0.042
5 — — 0.389 0.234
6 0.455 0.031 0.142 0.340
5 3 — 0.191
— —
4 — 0.319 — 0.191
5 —. 0.093
— —
—
6 — 0.182 0.167
7 — — 0.360
—
6 4 0.315 0.098 0.021 0.235
5 — — 0.154 0.141
6 0.255 0.364 0.364 0.130
7 — — 0.347 0.124
8 0.431 0.039 0.116 0.371
*The entries are the squares of the Clebsch-Gordon coefficients for each indicated transition. Thus, relative
B(E2) values connecting states /., Jf in bands with K., Kf are B(E2 JX.-tJfa) ~ (J.K2&K l^)2.

involved in these branching ratios. However, transitions involving K = 0 and

K - 2 bands are so important and so common that it is useful to collect the
results here. Table 6.6 shows the relative B(E2) values for transitions involv-
ing low-spin states in K = 0 and K = 2 bands.
It is interesting to compare them with the data for deformed nuclei. To this
end, we show comparisons for three nuclei in Table 6.7, of which one, 154Sm, is
situated at the beginning of the deformed region while the other two, 168Er and
178
Hf, are near midshell. The table shows a number of very interesting
features:
1. The general agreement is remarkably good, indicating that these simple
expressions are a reasonable leading-order approximation. Despite the
deviations to be discussed at considerable length next, it is important to
stress that the approximate validity of the Alaga rules is one of the
strongest arguments for axially deformed nuclei and for the concept of
separable rotational motion. Nevertheless, there are substantial devia-
174 Collectivity, Phase Transitions, Deformation

Table 6.7.Comparison of some relative B(E2: y—»g) values in deformed rare earth nuclei with the
Alaga rules*

Relative B(E2:JJf)
U4 168 178
J. Jf Alaga Gd Er Hf

2 0 70 43 54 88
2 100 100 100 100
4 5 14 6.8 5.8
3 2 100 100 100 100
4 40 105 65 52
4 2 34 16 20 18
4 100 100 100 100
6 9 14 —
—
—
5 4 100 100 100
6 57 — 123 107
6 4 27 — 12 18
6 100 — 100 100
8 11 — 37 —
"One transition is normalized to 100 for each initial state. The Alaga rule entries are relative values from Table
6.6 for the K = 1 -> K = 0 case.

tions from them and their study greatly deepens our understanding of
deformed nuclei.
2. The deviations increase substantially with increasing spin.
3. Transitions in which the spin increases (Je > /) are nearly always empiri-
cally larger than the Alaga rules, while spin decreasing transitions (Jf<J)
are nearly always smaller.
4. The deviations can become quite large, leading to factors of three or four
discrepancies from the predictions.
5. The deviations are, on average, larger in Sm than in Er and larger in Er
than in Hf.
Combined with all the evidence from rotational energy sequences, meas-
urements of quadrupole moments, and the like, point 1 provides a vast body of
evidence that supports the idea of a superposition of rotational and intrinsic
motion and the approximate separability of the two. Point 5 suggests that this
separability is most applicable in midshell and least just after the transition
regions from spherical to deformed nuclei. This is reasonable, of course, since
the energy scale of rotational motion decreases systematically toward mid-
shell, and therefore, the distinction in energy between rotational and vibra-
tional behavior is larger there than closer to the vibrational regions at the
beginning and end of major shells. We shall soon see more dramatic evidence
of this point shortly in terms of a systematic measure of the rotation-vibration
coupling. First, however, in order to understand points 2, 3, and 4, and in
particular why point 4 does not indicate a serious breakdown of the rotational
description, we must introduce the concept of bandmixing and discuss a
quantitative formalism to treat it in a simple way.
 Collective Excitations in Even-Even Nuclei 175

Bandmixing and Rotation-Vibration Coupling

Bandmixing is a widespread phenomenon in even-even nuclei. We shall limit
our discussion to its simplest and most common manifestations, namely A/C = 2
mixing between the 7 and the ground bands and AK = 0 mixing between /J and
ground bands. This is not to minimize the importance of /J-y mixing or the
mixing of two-quasi-particle intrinsic excitations with vibrational states: the
same physical concepts and formalism also apply to such cases but the extent
and systematic behavior of y-g mixing makes it the most interesting and
informative to study. The basic scheme is simply an example of two-state
mixing. For y-g mixing, the ground and 7 band wave functions can be written
as

where e' is the small mixing amplitude of each band in the other. It is
convenient to separate the spin dependent and spin independent parts of the
mixing by writing

In order to derive the spin dependence, we need to anticipate a result from

Chapter 8 on Coriolis mixing in the Nilsson model. Coriolis mixing is a well-
known effect in any rotating system and arises from the transition from a body-
fixed (nuclear) frame of reference to the laboratory. We will show in Chapter
8 that the Coriolis effect in nuclei mixes intrinsic states differing by &.K = ±1.
The dependence of the Coriolis mixing matrix element on the total angular
momentum J is contained in the mixing operator J+, given by (see Eq. 8.2)

where K is the lower AT value. We now interpret y-> g mixing as proceeding via
a two-step Coriolis effect through an intermediate K = 1 band (which need not,
and generally is not, known empirically). For weak mixing, this can be viewed
as a sequence of two separate two-step mixing effects.
We know from the discussion in Chapter 1 of weak two-state mixing that if
three states 0,, 02, 0, mutually mix, the mixing of states 02 and 03 gives

Then, if the already mixed state 02' mixes with state 0P we have

or, since a and

Thus, the overall mixing amplitude of state 03 in state ^ is simply given by the
product of the individual two-state mixing amplitudes J312 and /323.
176 Collectivity, Phase Transitions, Deformation

Applying this to the present case, we have the mixing sequence (K = 0) ->
(K = 1) -»(K = 2). Hence the spin dependence of the AAT = 2 mixing amplitude,
/XO is

Similarly, for mixing between the j8 (K = 0) and ground bands, the mixing
sequence is (K = 0) -> (K = 1) -»(K = 0). Then/),(./) is given by

Note that bothf^f) and/r(/) -» /* for large J: the band mixing increases rapidly
for high spin.
The spin dependence of e' explains point 2 concerning the increase of the
deviations from the Alaga rules with increasing spin. However, we have yet to
explain why these deviations can be so large without implying a corresponding
destruction of the entire rotational picture on which the Alaga rules and the
present formalism are based.
We can now calculate the interband E2 matrix elements very simply using
the admixed wave functions of Eq. 6.22.

The first term in Eq. 6.27 is the direct matrix element in the absence of
mixing. Thus, the perturbed matrix element can be written as a sum of a direct
term plus a contribution proportional to e'. In deriving this expression we
have dropped terms in e' 2 since the mixing is assumed to be small. Each of the
two terms multiplying er" is proportional to a Clebsch-Gordon coefficient
multiplied by the intrinsic quadrupole moment of the 7 or ground band.
Therefore, even if we assume these intrinsic moments to be equal (as is
commonly done since the deformation does not differ much from band to
band), the K dependence of the Clebsch-Gordon coefficients prevents this
term from vanishing.
In the case of /} -> g mixing, exactly the same formalism applies with a
substitution of ff for/ One interesting result for the special case of transitions
that do not change spin (/. = / ) follows immediately. For identical quadrupole
moments, the two terms multiplying e' are identical and vanish: The /? —» g
bandmixing has no effect on transitions which do not change the spin. This is
a special case of the result derived in Eq. 1.17.
Incorporating the spin dependence of € and expressions for the Clebsch-
Gordon coefficients in Eq 6.27 leads to a general form for the effect of band-
mixing on interband B(E2) values. We obtain in this way the well-known ex-
pressions:
 Collective Excitations in Even-Even Nuclei 177

Table 6.8. Correction factors F (7., J.) and F^ (7., 7) for y-> g and ft -> g reduced E2 matrix
elements due to y- g and /3 - g bandmixing
7. ] Correction factor
—— j-—

7 1 + (27 + 1 ) 2 1 + 2 ( 2 7 - 1 ) 2 ,
/-I / ! + (/,+2)2 -
/ I, l+2Z r 1
X+i f, i-(J}-i)zr -
7,+ 2 7) l-(27j+l)2 y 1-2(27,+3)2,
*Riedinger, 1969.

Here 50(E2) is the unperturbed value and Z and Z^ are bandmixing

parameters proportional er and ep, respectively. The functions T7^,./,) and
/"".(/p, /g) are given in Table 6.8 for the three possible cases of AJ = 0, ±1, ±2.
Clearly the case A7 = ±1 does not apply to the ft -> g transitions, and the result
obtained earlier for A/ = 0, j3 —»g transitions is reflected by the value of unity
for/?,(/, = 7,)
One can analyze experimental data in terms of this formalism in several
different ways. One is to extract a Z value from each branching ratio between
a pair of bands and then test for a consistent value. An example of this is shown
in Table 6.9, where the Z values for 152Sm are given. Clearly in this case,
Z r ~ 0.078 provides good agreement with the data.
Another approach exploits a particularly useful form of Eq. 6.27. We will
give it in a form that is applicable to both /3 -» g and 7 -> g transitions by
combining the spin dependence of the fft or /y functions with that of the
Clebsch-Gordon coefficients. We obtain

For y-> g transitions Ml and M2 are defined by

Table 6.9. Z values for 152Sm*

Branching Experiment Alaga Z
Ratios (xlO2)
4
r->4«
^—^ 11.2(19) 2.94 8.1(8)
3^4g
— 1.00(5) 4 7.7(5)

^~ 10.0(14) 20 6.7(18)
2 2
r-* g
— 2.38(18) 1.43 8.8(14)
lyr-tOg

~——
z >4
r~ « 4.16(61) 14.0 7.6(1.1)

*Errors given on last digit. Riedinger, 1969.

178 Collectivity, Phase Transitions, Deformation

and are related to Z by

For (5 -> /transitions, Ml = (<t>p I A/(E2) | 0^) and MI = (5l\6n)mQ0ep

Thus, Mj is essentially the direct intrinsic AX" = 2 matrix element (the
correction term, -4M2 for 7 - g mixing, is normally very small) and M2 is
proportional to the spin independent mixing amplitude e^. The advantage of
this form of Eq. 6.30 is that it can be rewritten as (taking the y—»g case to be
specific)

A plot of the left side against the spin function on the right is a straight line with
intercept Ml at J = / and slope M2. From such a plot, called a Mikhailov plot,
one can extract directly from the empirical results both the direct intrinsic
unperturbed AK - 2 matrix element and the mixing amplitude e (<^M2), pro-
vided the data can be fit by a straight line. Deviations from a straight line can
arise from several sources: unequal quadrupole moments of the bands, more
than two-bandmixing, undetected Ml components in the interband transi-
tions, or two-bandmixing that follows a different spin dependence than that
given by the fp(J) and/(7) functions.
An example of a Mikhailov plot, for 168Er, is shown in Fig. 6.14 (the dashed
line labeled IB A will be discussed later). It is clear that the data points are very
well approximated by a straight line, thus validating the use of the Mikhailov
formalism. Such data are typical of deformed nuclei.
The use of the bandmixing formalism in either the Zror Mikhailov forms is
a powerful tool for analyzing deviations of relative E2 transitions from the
rotational (Alaga) values and for studying rotation-vibration interactions.
Empirically, it is invariably found that M2« Ml and negative (by convention,
Ml is positive): therefore Z is positive. In this way, Z values have been
extracted for a number of rare earth nuclei. The results are summarized in Fig.
6.15.
These systematics exhibit a parabolic behavior that minimizes at midshell.
This smooth pattern highlights, indirectly, the collective structure of the unper-
turbed states involved: if the mixing were with single-particle excitations, it
would surely be more erratic. As we shall see in Chapter 9, however, the
microscopic structure of collective vibrations changes smoothly and slowly
with N and Z. The minimum in Z at midshell, and the generally small values
 Collective Excitations in Even-Even Nuclei 179

Fig. 6.14. Mikhailov plot for y-> g transitions in 168Er (Warner, 1982).

Fig. 6.15. Empirical systematics of Z^in the rare earth region (Casten, 1983). N is half the number
of valence nucleons.
180 Collectivity, Phase Transitions, Deformation

of Z , demonstrate the overall validity of the separation of rotational and

vibrational degrees of freedom and show that, as expected, this separability is
best at midshell.
With this bandmixing formalism and the empirical results in hand, we can
now address points 2, 3 and 4 mentioned earlier. First of all, Eq. 6.32 shows
that, regardless of the sign of M2, the effect of bandmixing (on A/ * 0 transi-
tions at least) must increase with increasing spin as we anticipated earlier from
the spin dependence of the mixing amplitudes e and £„. Second, the negative
values of M2, that is, the positive slopes in a Mikhailov plot, and the fact that
the abscissa is positive iorJf>J., implies that the B(E2) values are increased for
spin-increasing transitions and decreased f or spin-decreasing transitions, as we
observed in the examples given in Table 6.7.
This point is also clear from inspection of the analytic formulas in Table 6.8.
Since M2 has the opposite sign to Z , the data in Figs. 6.14 and similar data for
other deformed regions show that Zr is always positive. Then, we see from
Table 6.8 that, for spin increasing transitions, the B(E2) correction factor
always has the form

while for spin decreasing transitions, we have

where g(J) and h(J) are positive functions of the final spin /.
It is worth working through an explicit example to see how the bandmixing
technique is used. We take the case of y-ground mixing in 168Er and use both
the analytic approach with Table 6.8 and the Mikhailov formalism.
According to Tables 6.6 and 6.8,

where the first factor is the unperturbed (Alaga) ratio. From the experimental
value of 1.85, we obtain Z = 0.044. Similar values are obtained from other
transitions. A good average value is Z ~ 0.038. The small magnitude of Zr
confirms the adequacy of the two-state bandmixing approach.
Turning to the Mikhailov approach, which generally is easier and yields
interesting physics more directly, Fig. 6.14 gives

Thus, we can immediately deduce the direct, unperturbed /-> g E2 matrix

element

Using Qg = 7.61 eb, the spin independent part of the mixing amplitude is
 Collective Excitations in Even-Even Nuclei 181

The full mixing amplitudes e'(J) = VI e f(J) are then -0.0053 (2+),
-0.012 (3+), and -0.021 (4+).
We can now calculate the actual mixing matrix element since for such small
mixing we have, from Eq. 1.10

Neglecting the tiny difference between perturbed and unperturbed spac-

ings, and taking E^^-E^$ - 741 keV we get

In a way, it is remarkable how much detailed information, including a

mixing amplitude, an interaction matrix element, and even an absolute unper-
turbed transition matrix element, can be obtained in this simple way and only
from the measurement of relative interband B(E2) values, without measuring
absolute transition rates.
Finally, we now see that the rather large changes in the interband B(E2)
values result from extremely small residual interactions, on the order of a few
keV, and mixing amplitudes ~10~2-10~3. Returning to our earlier point 4, we
see that the rotational description of the wave functions is still an excellent
approximation although certain observables deviate substantially from their
rotational predictions. Later we shall see other cases, such as Coriolis mixing,
where small disturbances of pure wave functions grossly affect certain observ-
ables and, conversely, where the measurement of those observables provides
very sensitive probes of specific wave function components.
The fundamental reason that small interactions such as the one we are
considering can lead to such large effects is obvious from Eq. 6.27: the mixing
introduces an effectively intraband contribution to the originally interband
transition. The 4 + —> 2 + transition, for example, contains small amplitudes for
the very large 4 + -> 2* and 4g+ ->2?+ rotational matrix elements. From this, we
can immediately appreciate the well-known empirical fact that contrary to
interband transitions, intraband transitions are virtually unaffected by band-
mixing because the effect is reversed, namely, adding a small interband ampli-
tude to a much larger intraband amplitude.
By formalizing this argument we can deduce some interesting results. We
consider the set of bands shown in Fig. 6.16, where we illustrate states of spin
/,/+ !,/ + 2 occurring somewhere in these bands. We donor restrict ourselves
to small mixing. First, let us isolate the two-band system of intrinsic states 1
and 2 and actually calculate the relevant B(E2) value for the transition
between states of spin / and J. If the mixing is large, we might expect
transitions between bands to be comparable to those within a band. However,
we shall see that under one rather reasonable assumption, this is not the case.
We explicitly write the B(E2) value, using an obvious notation for the initial
and final wave functions analogous to the notation in Eq. 1.7 except that it
distinguishes amplitudes a, a7, ft, ft' for the spins J and /. We have, for an
intraband transition in band 2,
182 Collectivity, Phase Transitions, Deformation

Fig. 6.16. Set of admixed bands (see text).

But, here, the interband matrix elements connecting unperturbed states in

bands 1 and 2 are negligible compared to the intraband rotational matrix
elements. We assume for simplicity that the intrinsic matrix elements are band
independent and obtain

If we now assume that the mixing interaction, though possibly large, is not
very spin dependent, then the composition of the mixed wave functions will
also not depend much on spin and therefore, a~ a 'and fi~ fi'. But, then, the
factor (aa' + flfi') ~ 1 by orthonormality and the intraband transition has an
identical B(E2) value as in the unmixed case.
Extending this argument to multiband mixing, the factor (aa' + /J/T) in Eq.
6.37 will simply be replaced by (aa' + ftp' + 77' + 88' +...). If, again, the
primed and unprimed mixing amplitudes are approximately equal, this is just
the orthogonality sum, which is unity. (For interband transitions, the ampli-
tude sum is (0,0,' + /3j/32' + ...), and in this case nearly vanishes by the same
orthogonality argument.) Thus we see that, although the mixing is large,
z'm/'aband transitions are barely affected and retain their normal rotational
strengths.
This has many repercussions, two of which are worth citing briefly. It
 Collective Excitations in Even-Even Nuclei 183

means, for example, as we argued already, that Alaga rules for intraband
transitions are essentially unaffected by mixing. Thus, observed deviations
from the Alaga rules can be ascribed to other mechanisms (e.g., Ml compo-
nents) and can be used to estimate these. Second, consider heavy ion reactions
that bring large amounts of angular momentum into the nucleus, which then
decays by a series of cascade transitions (see Chapter 10). It has been observed
that these cascades flow through many rotational bands, but that the popula-
tion within a band tends to remain intact as J decreases, even though these
relatively high-lying quasi-particle excitations are expected to mix considera-
bly. The preceding derivation provides a simple explanation: the mixing can
indeed be strong, but as long as it does not change rapidly with J, the z'mraband
transitions are only slightly affected and remain dominant.
In closing this section, we note that extensions of the formalism to include
j9 - 7 mixing have also been developed and are available in the literature. One
point that will be useful in our later discussion of the IBA can be deduced
immediately without a formal development of the mixing expressions. The
effects of/?- 7 bandmixingon ft —>gand 7—> g transitions are second order
and generally weak; however, since j3 -> 7 transitions are forbidden in the
absence of mixing, such mixing can strongly break this fundamental selection
rule. The expression for /} —> 7 transitions in the presence of j3 - 7 mixing is
analogous to those we generated for the ft — » g and 7—> g cases, except that
there is no longer a direct term and hence the entire transition strength arises
solely from a mixing term proportional to Z . Thus, in ratios of /3 —» 7
transitions, Z. cancels out. Therefore, although the finite transition matrix
elements arise from mixing, branching ratios are independent of the strength of
that mixing and are given only by ratios of functions of 7, and 7 . We note for
future reference that in the IBA model, ft -> 7 transitions are, in contrast,
allowed for deformed nuclei but their branching ratios depend on the detailed
structure (in effect on the value of the asymmetry parameter 7). We will
discuss this further later in this chapter.
Having discussed the low-lying, intrinsic excitations of axially symmetric
nuclei, we can return to the question of rotational energies and corrections to
the simple first-order expressions in Eqs. 6.13 and 6.14. It was useful to discuss
these intrinsic excitations, in particular, the bandmixing between them, first,
because the corrections to the symmetric top formula are intimately connected
with excursions from axial symmetry and rotation-vibration coupling. Indeed,
the first order rotational expression makes several implicit assumptions, the
most important of which are that there is no coupling between rotational and
intrinsic degrees of freedom and that /? is independent of J. These two
assumptions are, in fact, related. As the nucleus rotates, it experiences a
centrifugal force that tends to increase the deformation and moment of inertia
and decrease the rotational spacings, and leads to an enhanced coupling to
vibrational modes (recall Eqs. 6.25 and 6.26, which show that this increases
with spin). There are several ways of incorporating these effects into a
rotational energy expression. One of the first and most common is simply to
expand the rotational energy in powers of 7(7 + 1) and keep the second term.
184 Collectivity, Phase Transitions, Deformation

Fig. 6.17. Empirical ground band levels of Yb compared with various models. The labels ab and
A, AB, ABC refer to the coefficients in Eq. 6.40 and in the expansion of rotational energies in
powers of J(J +1) (see Eq. 6.40 and following discussion).

One then has

where A - h 2/2I. (We will derive this formula in a moment.) From our earlier
comments, we know that empirical values of B are negative. If they are also
small (BIA « 1), the expansion converges rapidly and Eq. 6.38 will be a
significant improvement.
In some cases, still higher-order terms such as CJ3(J + I)3 are necessary to
produce adequate fits for higher J values. Rather than explore this, we shall
turn shortly to an alternate expression that automatically includes Eq. 6.38 and
all higher-order terms. First, we show an example in Fig. 6.17 of the ground
state rotational band of 168Yb compared with the energies calculated from Eqs.
6.13 and 6.38, as well as other expressions to be discussed. Evidently, the first-
order expression (Eq. 6.13) is reasonable only for very low-spin states. Equa-
tion 6.38 (AB in the figure) is an improvement for higher spins, although it too
encounters serious difficulties for still larger./. A fit with the CJ3(7 + I) 3 term
(ABC) included further improves the predictions, but is also inadequate for
large 7: the opposite signs empirically deduced for B and C tend to produce
wild oscillations in predicted energies (compressions of levels, even spin
inversions) at high enough./ values.
An alternate approach to incorporating rotation-vibration or centrifugal
 Collective Excitations in Even-Even Nuclei 185

effects into the rotational energy expression is to make the moment of inertia
spin dependent. This approach is known as the Variable Moment of Inertia
(VMI) model and has enjoyed considerable success. In general, its predictions
are better than those of Eq. 6.38, and it is not limited to the realm of strongly
deformed nuclei. Figure 6.17 includes VMI predictions and shows their
advantages. We shall not dwell on this approach, as it has been extensively
covered in other literature.
Interestingly, it is easy to see how both effects (a change in the moment of
inertia and the addition of a higher order term) result immediately from the
effects of f-g bandmixing.
We have seen that the mixing is generally small so we can use the approxi-
mation of Eqs. 1.12 to write the energy shift (lowering) of the ground state
band as A£ *sb(J) = V2/AE2 . But, from Eq. 1.10, the mixing amplitude
er' =V2e/r(y) = VIAE^. So, AEf = 2 eff(J). Hence, from Eq. 6.25,

The second term is the promised correction to the standard rotational formula,
and can give the variation with / of the inertial parameter ft 2/27, while the first
gives the required second-order correction term. From this derivation it is
clear that Eq. 6.38 is, as we implied earlier, ultimately connected with the
concept of rotation-vibration coupling (bandmixing) and also that it implicitly
assumes small mixing.
When / becomes large enough such that er/r(J) ~ 1 we must anticipate a
breakdown of Eq. 6.38 and thus a need for many higher-order terms or an
alternate formula. We have seen this effect empirically in the failure of Eq.
6.38 for J> 14 in Fig. 6.17.
However, there is a much superior rotational expression that is valid for
even higher spins that unfortunately has not been discussed much in the
literature. It automatically gives Eqs. 6.13 and 6.38 as limiting cases, automati-
cally includes all the higher order correction terms, and moreover, contains a
specific relationship between the coefficients of each successive term. One
simply writes the two-parameter formula

where a and b are parameters.

This expression can be derived in the Bohr-Mottelson picture by including
small deviations from axial symmetry. A trivial rationale for this was pre-
sented by Lipas many years ago. Suppose that we make the ansatz that we can
write the moment of inertia / as a function of excitation energy:
186 Collectivity, Phase Transitions, Deformation

Then, substituting this into Eq. 6.13 gives

or

Hence

or, taking positive energies,

which is simply Eq. 6.40 with a = a/2/3 and b = 2h 2fi/a2. Note that energy ratios,
such as Ejl E2\, depend only on the single parameter b in Eq. 6.40. Neverthe-
less, this formula is far more accurate than any of the expressions we have con-
sidered, as shown in Fig. 6.17 for 168Yb where the predictions are compared
with one-, two-, and three-term expansions in / (J + 1) and with the VMI
model. Its success extends to softer (transitional) nuclei (e.g., 152Sm, 184Pt).
Since the expression works so well for higher/, we anticipate a later discussion
to caution that it is only applicable below any "backbend" that may be present.
Aside from its empirical success, Eq. 6.40 is interesting because, for rela-
tively low spins such that bJ(J + !)«!, expansion of the square root naturally
recovers the second (and higher) order terms in the rotational formula of Eq.
6.38:

Here, however, the coefficients of each power of J(J + 1) are interelated,

whereas in Eq. 6.38, they are arbitrary parameters. It is remarkable that this
constrained version of the expansion in powers of the angular momentum
produces such an excellent fit.
For large J such that b.l(J + 1) »1 (this typically requires J ~ 30), this
expression is almost linear in /, reflecting the enormous compression of the
ground band due to both bandmixing effects and to centrifugal stretching.
Thus, Eq. 6.40 incorporates the limit of small mixing (Eq. 6.38), but also
extends into spin regions where the mixing is large.

6.5 Axially Asymmetric Nuclei

The models we have discussed thus far incorporate excursions from axial
symmetry that are both small and dynamic. Certainly, such an approach
accounts reasonably well for the deviations of most well-deformed nuclei from
the properties of the pure axial rotor. However, there have long been indica-
tions that larger and possibly permanent (static) asymmetries also occur.
 Collective Excitations in Even-Even Nuclei 187

Naturally, this would lead to more radical departures from the energy and
transition rate expressions we have considered. In fact, in certain limiting
cases of large asymmetry, new selection rules appear. In another sense,
however, such models for larger asymmetry are extensions of the small excur-
sions from axiality dealt with so far, and their predictions go over into the latter
as y-» 0°. It also turns out that many predictions of models for large, fixed
asymmetries y are identical, or nearly so, to models incorporating dynamic
fluctuations in y so long as %. d in one equals yms in the other.
The best known model of fixed stable asymmetry (triaxiality) is that of
Davydov and co-workers developed around 1960. Here, the potential V (y) is
envisioned to have a steep, deep minimum at a particular value of y so that the
nucleus takes on a rigid shape with that asymmetry.
We have seen that, if the rotational and vibrational motions are not com-
pletely decoupled, and there is an interaction (mixing) between the y and
ground bands, the latter will acquire a finite yrms and K will no longer be a good
quantum number. Therefore, it is not surprising that in the Davydov model K
is not a good quantum number either. Here, however, since y can be large, the
K admixtures can reach levels far beyond those we have encountered.
The relation between the Davydov model and models with axially symmet-
ric but ysoft potentials runs deeper than this. In a nucleus with such a
potential, the greater the softness the lower the y vibration will lie, and the
larger ymt will be in the ground state. In the Davydov model there is no
distinction in intrinsic structure between what is normally called the ground
state rotational band and the y vibrational states. The levels of these two
bands simply become the so-called normal and anomalous levels of a new
ground state band whose energies depend explicitly on y, which can take on
values fromO 0 —> 30° (prolate symmetric —»maximum asymmetry: 30° < y< 60°
corresponds to the "oblate" region of asymmetry). Figure 6.18 shows the
lowest levels as a function of y and clearly illustrates the descent of the y
vibrational levels. Indeed, for y> 25°, E2+<E4+. In contrast, as y-» 0°, the
normal levels 0+t, 2+v 4+v 6+p which are rather insensitive to y, go over into
those of an axially symmetric ground state band, while the "y band" energies
increase rapidly. An important feature of the anomalous levels is their energy
"staggering": They tend to be grouped into couplets as (2+, 3+), (4+, 5+)...
The behavior in Fig. 6.18 is easy to understand. As y increases, the nucleus
becomes increasingly flattened. Therefore, states whose wave functions are
predominantly aligned in the direction of the flattening attain lower energies,
since the nuclear force is attractive and they are, on average, closer to the bulk
of the nuclear matter. This is exactly the case for the K - 2 (and higher K)
levels which, therefore, rapidly decrease in energy with increasing y. They also
mix with the normal ground state band levels (yrast states) and, as we have
seen, K is no longer a good quantum number. The leveling off of the energies
of the anomalous levels for y >25° is easily understood in terms of that mixing.
In Table 6.10 we give a number of interesting quantities relating to the
Davydov model, including the amplitude for K = 0 in the 4^ and 42+ states as a
function of y. For y between 25° and 30°, the major amplitudes in the 4 + state
188 Collectivity, Phase Transitions, Deformation

Fig. 6.18. Normal and anomalous levels of the triaxial rotor (Preston, 1975).

actually interchange so that the K = 0 amplitude is larger than the K = 2 one.

(The 4j + amplitudes do not quite "cross" since there is actually substantial
three-state mixing involving the 43+ state.) The interchange of amplitudes in
the 42+ level occurs near the energy inflection point in Fig. 6.18. The decreasing
trend of the quasi-y-band, or anomalous level energies, would have caused
these energies to cross the normal levels at this point. Instead, the interaction
causes a repulsion. This is a nice example of this two-state mixing effect
discussed in Chapter 1.
In practical applications of the Davydov model, one usually extracts 7 from
the energy ratio E2^l E?.\ of the first two 2 + states. This ratio is given in Table
6.10 for several values of 7 and is plotted in Fig. 6.19 (along with two B(E2)
ratios). It can be calculated for any 7 value from the expression
 Collective Excitations in Even-Even Nuclei 189

Table 6.10. Some useful predictions of the asymmetric rotor (Davydov) model*

r^> 0° 5° 10° 15° 20° 25° 27.5° 30°

£
E2V 2+1 64.2 15.9 6.85 3.73 2.41 2.10 2.00
B('E2:2+2->2+i
1.43 1.49 1.70 2.70 5.35 20.6 82
%(E2:2+2-*tfi

K=0 1 1 0.999 0.993 0.955 0.852 0.792 0.739

^ K=2 0 0.003 0.030 0.114 0.296 0.522 0.605 0.661
K=0 0 -0.003 -0.030 -0.114 -0.296 -0.523 -0.602 0.559
¥
^ K=2 1 1 0.999 0.993 0.954 0.842 0.754 0.500

E(EZ2\-><fi 1 0.993 0.972 0.947 0.933 0.955 0.985 1

+
B(E2:2 2->0 i +
0 0.0074 0.028 0.053 0.067 0.425 0015 0
BfE2:2+2->24i 0 0.011 0.051 0.143 0.357 0.865 1.23 1.43
BfE2:2+2^>oV
: 0 0.0074 0.0288 0.056 0.072 0.0445 0.015 0
BfE2:24i ->0+i
BfE2:4+2-*2fi ) 0 0.0004 0.011 0.008 0.0004 0.021 0.018 0
Bf E2:4 i ->2<-2 )
+
0 0 0.0023 0.010 0.033 0.039 0.016 0
BfE2:4+2->4Ti ) 0 0.0138 0.0624 0.167 0.313 0.311 0.271 0.273
Bf E24+1 -^2+i | 1.429 1.418 1.395 1.377 1.372 1.365 1 378 1.389
>
Bj E2:3+i -><?\ ) 0 0.006 0.034 0.130 0.406 0.821 0.955 1
4 +
Bf E2:3 i ->2 i ) 0 0.0132 0.0492 0.095 0.12 0.079 0.027 0
B(E2:3+i-»2Y 1.78 1.77 1.74 1.69 1.67 1.70 1.76 1.78

*B(E2) values are in units of e Q 0/16ff.

where

The individual energies are equal to the respective numerators and denomina-
tors multiplied by (9/sin23y). Note that X-> 1 for y-»0°. Thus, £ 2 £/£ 2 ;->°°
for y-> 0° as seen in the Table and figure. Also, X -> 1/3 for y-> 30°. Table 6.11
gives the empirical values of E^ I Ei\ for a number of heavy rare earth nuclei
as well as the associated y values. If, on account of the anomalously low
E^*^IE^\ ratios, these nuclei are considered to have large rigid triaxiality (this
is the term usually used for the concept of fixed asymmetries), then these y
values are the only input needed to make Davydov model predictions for other
observables. Such predictions are included in Table 6.10 for several y values.
The branching ratio,

is also a useful indicator, as shown in the same table and in Fig. 6.19. R2 can be
written analytically and calculated for any y from
190 Collectivity, Phase Transitions, Deformation

Fig. 6.19. Dependence of several observables on ^(compare Fig. 6.42).

where the numerator and the denominator are the individual B(E2) values.
Note that both B(E2) values in R2 vanish for 7= 0°, yet they have a finite ratio
that is the Alaga rule: the vanishing is reasonable since for 7= 0°, E2$ I Ei \ -» °° ,
corresponding to infinite rigidity in the 7direction and to vanishing vibrational
amplitude. R2 increases rapidly with 7 and R^ —> °° for 7= 30°. This latter result
is identical to the selection rule for an alternate model of axial asymmetry that
we will soon discuss, the 7 flat or y-unstable model of Wilets and Jean in which
the 2 transition is forbidden.
Some other B(E2) values and branching ratios are given in Table 6.10.
Those corresponding to rotational transitions in either the 7 or ground bands,
are nearly 7 independent. Others vanish at both 7= 0° and 30° but attain small,
 Collective Excitations in Even-Even Nuclei 191

Table 6.11. Values of E2+/E2+ for some deformed and transitional nuclei and the corresponding
y values (rounded to nearest degree)
Nucleus E 7
^'EA
152
Sm 8.9 13°
160
Dy 11.1 12°
168
Er 10.3 13°
172
Yb 18.6 9°
176
Hf 15.2 10°
182-^y 12.2 12°
184
0s 7.9 14°
188
0s 4.1 19°
192
0s 2.4 25°
196p, 30°
1.94

finite values for intermediate 7 values. These are transitions between normal
and anomalous levels. The 22+ —> 2,+ and 3,+ —> 4^ transitions form a third
category: small at 7 = 0° and rising rapidly toward 7= 30°. This behavior is
easily understandable if we note that, in the 7= 30° case, the Davydov model
has the same selection rules as both the quadrupole vibrator model and the
Wilets-Jean 7-unstable model. For example, at 7 = 0°, the 22+ -> 2^ and
3a+ —> 4j+ transitions are interband (7—> g): given the built-in stiffness in 7, they
must be forbidden. At 7= 30°, the 3:+ state is analogous to a three-phonon
level, the 4^ level to a two-phonon excitation, so the transition becomes
allowed. Similarly the 22" —» 2 T + transition is analogous to a 2 —> 1-phonon
transition.
Since finite 7 values correspond to mixed K values, one might expect a close
relation between B(E2) values for finite 7 in the Davydov model and in the
bandmixing formalism. This is indeed so. Consider, as an example, the case we
worked out earlier of 168Er. We found that the Mikhailov plot analysis gave a
full mixing amplitude in the 4+ states of e '(4*) - 0.021. In the Davydov model,
the experimental ratio R^ = 1.85 yields 7- 11° (see Table 6.10). This, in turn,
corresponds to a y-» g mixing amplitude of 0.03, which is quite close to the
bandmixing result. The agreement is not exact since the comparison is not
quite on an equal footing. In the bandmixing case, the spin independent
mixing amplitude was deduced from a Mikhailov plot, which gives an overall
average value for all transitions, while the Davydov mixing value was deduced
from R2 alone. In any case the essential point is that both the bandmixing
formalism and the Davydov model lead to K mixtures in 7 and ground bands,
and give comparable mixing amplitudes and B(E2) values for small 7. (The
bandmixing formalism is a first-order perturbation treatment and is therefore
inapplicable for large 7.) Though the physical pictures are different, predic-
tions for many observables are nearly identical.
When we turn to the large 7extreme, one might think that the extremely
low-lying 7 band levels of the Davydov model would be an appropriate
signature for stable axially asymmetric shapes. However, we can also envision
large 7 values as dynamic quantities by picturing a deformed nucleus that is
192 Collectivity, Phase Transitions, Deformation

totally free to vibrate in the 7 degree of freedom. Such a nucleus is completely

7-soft: it is a "/-unstable" rotor. This corresponds to a nuclear potential
centered at a finite /3 but completely flat in the 7 degree of freedom. The
nucleus oscillates smoothly from 7= 0° - 60° and has yms = 30°. In the extreme
limit of complete 7 instability, known as the Wilets-Jean model, the "rota-
tional" energies are given by

where -d is a constant analogous to hz/2l and the levels are now classified
according to the quantum number A. This classification scheme is given in Fig.
6.20. The yrast levels have / = 2A. Note that each A value (for A > 1)
corresponds to more than one level and that the A values 2, 3, 4, ... include a
low-lying set of levels analogous to the 7 vibrational band and to the anoma-
lous levels of the Davydov model for large 7.
Now that we have the three basic extreme geometric models, the harmonic
vibrator, the axially symmetric deformed rotor, and the 7soft axially asymmet-
ric deformed rotor, it is interesting to compare their rotational energies. The
results are shown in Fig. 6.21. The spacing of the normal rotational levels of
the Wilets-Jean rotor are quite different than in the symmetric rotor. This is
easy to see if one notes that for these levels (the yrast states) A = J/2, so that
Ew^(/-yrast) = (-Q/4)J(J + 6), which increases with / considerably slower than
the /(/+ 1) law for a symmetric rotor. For example, Et\lEi\ = 2.5 compared
to 3.33 for the rotor and E6IE = 4.5 instead of 7. We note for later reference

Fig. 6.20. Ground band levels in the y-unstable or Wilets-Jean model. (Left) In terms of A
multiplets; (Right) Displayed in analogy to the quasi-band structures of a normal rotor.
 Collective Excitations in Even-Even Nuclei 193

Fig. 6.21. Dependence of ground band energies on spin for different models. An identical set of
curves is obtained in the U(5), O(6), and SU(3) symmetries of the IB A (see Eq. 6.72).

that the 7 unstable limit of Wilets-Jean is very closely related to the 0(6) limit
of the IB A Model.
Another characteristic of the Wilets-Jean model concerns E2 selection
rules. The allowed transitions must satisfy AA = ±1. Comparison with Table
6.10 shows that the Davydov model goes over to the same selection rules as the
Wilets-Jean model for 7= 30°, again highlighting the similarity of a y-flat model
with ymt = 30° to a rigid asymmetric model with fixed j- 30°. The Wilets-Jean
picture also resembles the phonon model in its E2 selection rule and gives
some identical predictions. For example, the 22+ -> 2/ transition
(A = 2 -^ A = 1) is allowed, as is 4^ -> 2+, but the crossover 2^ -»0^ transition
is forbidden (AA = 2), as is the 4t+ —> 22+(AA = 0). These are the same results
one would have if the O/, 2* and (4^, 22+) doublet were treated as the zero-,
one-, and two-phonon vibrational states. This is not surprising since the
potential for the spherical vibrator, while parabolic in ft, is independent of y, so
it is trivially 7-unstable as well. The real difference between the vibrator and
Wilets-Jean limits is that in one /3ave = 0, whereas the other has a deformed
minimum. Another difference between the two models (with the vibrator now
considered in its harmonic, degenerate multiple! limit) is that the vibrator has
a two-phonon triplet of levels 0+, 2% 4+ while the Wilets-Jean scheme has only
a 2+, 4+, doublet, the first excited 0+ state having A = 3. Thus, not only is it
higher in energy but its allowed decay is to the second2t level (2+2) rather than
the first as in the vibrator; more precisely, B(E2:02+ -> 220 /B(E2:02+ -» 2t+) -»°o
in the 7-unstable limit but is zero in the vibrator. The reader is warned,
however, that these differences do not necessarily persist in a sufficiently
anharmonic vibrator. We will encounter this point, and this close relationship
between vibrator and 7-unstable models, again, in our discussion of the U(5)
and O(6) symmetries of the IB A model later in this chapter.
Despite the obvious similarities between the Davydov and Wilets-Jean
models, there is one outstanding difference by which to distinguish them
194 Collectivity, Phase Transitions, Deformation

empirically: the energy staggering in the low-spin anomalous or y vibrational

levels is opposite in the two models. In the 7 unstable case, these levels group
according to (2+), (3+, 4+), (5% 6 + )... as seen in Fig. 6.20: in the simplest version
of this model, the levels within each couplet are degenerate.
It is interesting to use a quantitative measure of the asymmetry as a
distinguishing signature. To this end, we define the quantity AEj = E} - E} v
where the energies refer to the /-band, and then define the energy staggering,
ESV), as

For example,

These double energy differences involve three levels with spins J, J-l , and J-2,
and we use the convention that the level of spin J (the starting level) is always
of even spin. The usefulness of Es in comparing different models is evident if
we relate it to the energy of the first 2+ level, Etf. Although the harmonic
vibrator is not a rotational scheme, the levels of the "quasi-/ band" in that
scheme also display a staggering similar to the /-unstable model and include
the same degenerate couplets. We can therefore include this model in this
intercomparison as well. For the four cases of the symmetric rotor, the triaxial
rotor/ Davydov model with /= 30°, the 7-unstable or Wilets-Jean model, and
the harmonic vibrator, the following analytical expressions result:

(symmetric rotor)

(yrigid, 7= 30°)

(7unstable)

Generally, if Es(4) <Ez+/3 (7-soft case), the even spin states are depressed
relative to the odd spin; if £,(4)) > E^\l!> (7—rigid), the odd spin levels are
depressed relative to the even spin ones.
Before inspecting empirical values of Es(4), we note another interesting
feature of the vibrator limit. As we have seen, empirical values of the energy
ratio R = E2 h/Elph for vibrational nuclei are typically -2.2 rather than the strict
harmonic limit of 2.0. Table 6.4 gave the energy levels of the three-phonon
states, which include the 3+ and 4+ levels of the "quasi-7band" in terms of the
two-phonon levels. Es(4) cannot be defined analytically in this case since the
unknown energies of these 3+ and 4+ levels implicitly involve the energy of the
4^ level. However, if we make the approximation that the two-phonon 2^ and
4j + levels are degenerate (i.e., e2 = e4 in the notation of Table 6.4), then the 3+
and 4+ quasi-7band states will still be degenerate at an energy 3E2$ (R - 1).
 Collective Excitations in Even-Even Nuclei 195

Fig. 6.22. Systematics of Es(4)/£2t in the rare earth region. Predicted values for several models
are given at the right.

The energy anharmonicity of the two-phonon states is tripled in the three-

phonon states. Simple manipulations then give Es(4) = - E^\ (2R - 3). Of
course, in the harmonic limit 7? = 2, and this goes to Es(4) = - £2+ as in Eq. 6.48.
For the more typical case of R > 2, however, Es(4) becomes increasingly
negative compared to the harmonic value.
We see that Es nicely distinguishes different models ranging from vibra-
tional to rotational to axially asymmetric, and is also a useful signature of
transitional regions between these ideal limiting cases. We show in Fig. 6.22
the empirical Es(4) values for rare earth nuclei. Near the N = 82 closed shell,
the nuclei are close to vibrational with only a small component of /J softness.
Es(4) is negative and, in units of Ei\, ranges from -0.8 -> -0.4. As the neutron
number increases, the ft softness rapidly increases and Es(4) approaches the
rotational limit of + £2j73. The fact that most of the deformed rare earth
nuclei have Es(4) values slightly less than E2^/3 reflects both their predomi-
nantly axially symmetric rotational character and the presence of a small
amount of rotation-vibration coupling. Towards the end of the N = 82-126
shell, the behavior becomes more erratic, with the Pt isotopes displaying
tendencies toward extreme 7 softness.
Despite years of popularity of the Davydov model because of its simplicity
and analytic formulas, there is virtually no evidence for rigid triaxial behavior.
Axial asymmetry in nuclei seems associated with 7 softness instead.
In a later section of this chapter we shall see that such 7 softness also
 196 Collectivity, Phase Transitions, Deformation

characterizes the O(6) symmetry of the IBA and nuclei in O(6) -> rotor or
SU(3) transitional regions. To anticipate that discussion briefly, it is interest-
ing to note that the classic O(6) nuclei, 192~196Pt, have larger Es(4) values than
some of the other Pt isotopes, even though they are supposedly completely y-
soft. We can already guess from Fig. 6.22, however, that nuclei with nearly y
independent potentials but with shallow minima at 30° (that is, with the
addition of a small component of triaxiality at y= 30°) might have Es(4) values
higher than the extreme y-unstable limit. This is in fact the case, as will be
commented later in the discussion of the IBA.
The y-rigid-y-soft ambiguity is not the only one that can obfuscate an
interpretation of the structure. Figure 6.22 also shows the difficulty in distin-
guishing near harmonic vibrational structure from y softness. There is one
useful, albeit qualitative, indicator stemming from the systematic behavior of
certain absolute energies that can sometimes clarify whether a transitional
region is vibrational ~» symmetric rotor or axially asymmetric —> symmetric
rotor in character. This indicator is the relation between the energy of the 2 +
(normally 22+) level and the \+ state of the quasi-ground state band.
In a transition from vibrator to symmetric rotor, the 2 + level starts off
initially degenerate with the \+ level at rather high energies (since E 2 f is then
also rather high). As the phase transition develops, the 4,+ level drops rapidly
(as the 25+ energy drops) to become a member of the ground state rotational
band, while the 2* level remains rather high in energy (although it may drop
slightly). In a transition from an axial asymmetric rotor (either soft or rigid) to
symmetric rotor, the 2r+ and 4/ levels start out nearly degenerate (for large y)
and rather low-lying, but the former rises rapidly as y -»0° while the 4^ level
drops slightly as the deformation increases. These contrasting systematics are
shown for the Sm and Os nuclei in Fig. 6.23. This signature, while valid, cannot
be quantified as well as the energy staggering since the two levels involved
have intrinsically different structure in the rotational limit: the 4^ state be-
longs to the ground state band while the 2* state is an intrinsic excitation
whose energy depends on details of its microscopic collective wave function

Fig. 6.23. Systematics of low-lying levels in the Sm and Os nuclei. The symmetric rotor limit is at
the right in each case.
 Collective Excitations in Even-Even Nuclei 197

(see the discussion of the RPA in Chapter 9). In any case, it is evident that the
Ba-Gd region represents a vibrational-rotational transition, whereas the
Pt-Os region is axially asymmetric —> rotational.
This completes our brief survey of some of the essential features of geomet-
ric models for deformed even-even nuclei. These models have been
enormously successful, especially when the first order energy and B(E2)
predictions are modified by incorporating the higher order terms that reflect
changes in shape and adiabaticity of the rotational motion as J increases.
These models are all truly phenomenological in that they are applied to real
nuclei by inspecting empirical properties in order to assign the appropriate
shapes (J3,7, or higher-order moments). The models then provide a framework
for the prediction of numerous other observables.
In closing this section, it is important to stress that we have hardly exhausted
the collective vibrational modes in even-even nuclei. For example, we have
not touched on the most basic low-lying negative-parity excitations—octupole
vibrations—nor have we considered hexadecapole excitations. Just as quad-
rupole vibrations in deformed nuclei can have K = 0 or K = 2, octupole
excitations exist in K = 0,1,2, and 3 forms. In many deformed nuclei, several
of these have been identified, primarily via strong El and/or E3 transitions to
the ground state hand. The ordering of these octupole K values changes
systematically through a shell: near the beginning, the low K modes are lowest,
while they are highest toward the end. The basic physics, especially concerning
7-ray transitions into and out of these excitations, is similar in spirit to what we
have already outlined for positive parity levels. Though these excitations
account for most of the known negative parity states below the pairing gap in
deformed nuclei, they have been less well studied than their positive parity
quadrupole counterparts. Recently, however, they have taken on a new
interest in conjunction with, or rather as an alternative description to, the
possible existence of a-like cluster states, which are expected to have some of
the same properties. Hexadecapole vibrations are much less well known:
perhaps the most likely manifestation of their particular characteristics are in
the low-lying K = 4 bands in the Os isotopes.
We will not pursue a discussion of these higher multipole excitations here;
instead, we will return to the question of hexadecapole excitations and stable
hexadecapole deformations after our discussion of the Nilsson model in Chap-
ter 8, and to the structure and systematics of octupole excitations in Chapter 9
after our treatments of the TDA and RPA approaches to the microscopic
derivation of collective vibrations.

6.6 The Interacting Boson Model

Up until the mid-1970s the two principle strains of nuclear structure theory
were embodied in the shell model, which emphasized the single-particle
aspects of nuclear structure along with a careful accounting of the effects of the
Pauli principle, and the collective model pioneered by Bohr and Mottelson
and generalized by Gneuss and Greiner, Kumar, and numerous others. Each
198 Collectivity, Phase Transitions, Deformation

of these models, of course, had numerous offshoots: the shell model for
spherical nuclei led to the development of the Nilsson model for deformed
shapes and the collective models developed refinements that incorporated
axial asymmetry, higher moments in the shape such as hexadecapole compo-
nents, and the like. The link between these models was provided in the early
1960s by the development of microscopic approaches to collective motion
utilizing such techniques as the TDA and RPA, which we will discuss in
Chapter 9.
In 1974 the Interacting Boson Approximation (IB A) model was proposed,
which is based on a third approach that is group theoretical or algebraic and
that recalled methodology developed in the 1950s by Elliott and co-workers
for light deformed nuclei. The IBA has been extraordinarily successful, and
has also generated its own family of offshoots inspiring alternate, sometimes
competitive algebraic approaches such as the interesting pseudo-SU(S) and
symplectic group studies of Draayer and co-workers.
The IBA is a model for collective behavior. It has become customary to
refer to collective models of the Bohr-Mottelson type as "geometric" models
and those of the IBA or other group theory-based approaches as "algebraic"
models. Today one has a situation in which there is a triad of models—shell,
geometric, and algebraic—with which one can attack the basic problems of
nuclear structure. These models are not generally incompatible, although
there are differences in certain important details, but rather reflect three
approaches to nuclear structure that emphasize different complementary
aspects of that structure.
As we have commented repeatedly, the shell model rapidly becomes intrac-
table far from closed shells. In order to circumvent this, two basic alternatives
have been tried. In one, that of geometric models, the whole microscopic
approach is abandoned and replaced by a macroscopic one involving an
assumed or deduced nuclear shape, with rotations and vibrations about that
shape. The other, of which the IBA is an example, seeks to effectively truncate
the shell model space: the practical utility of such an approach depends on the
extent of the truncation, while its success depends on the appropriateness of the
truncation in isolating the key configurations involved (at least in the low-lying
states).
The truncation inherent in the IBA is shockingly extreme. For example, it
reduces the 3 x 1014 2* shell model basis states in 154Sm to 26! It is a wonder that
such a scheme can work at all, much less have the extensive and repeated
success it has enjoyed.
The basic idea of the IBA is to assume that the valence fermions couple in
pairs only to angular momenta 0 and 2 and that the low-lying collective
excitations of medium and heavy nuclei can be described in terms of the
energies and interactions of such pairs. These fermion pairs, having integer
spin, are treated as bosons (called 5 and d bosons for obvious reasons).
More formally, the model is founded on and embodies the following as-
sumptions and ideas:
• Closed shells of either protons or neutrons, and excitations out of them,
are neglected.
 Collective Excitations in Even-Even Nuclei 199

• The low-lying excitations of even-even nuclei depend only on the va-

lence space.
• The valence nucleons are treated in pairs, as s and d bosons, with angular
momenta 0 and 2. In the IB A-l no distinction is made between proton
and neutron bosons. The number of bosons is half the number of valence
protons and neutrons, both of which are always counted to the nearest
proton and neutron closed shells. This counting rule generates a finite,
fixed, number of s and rfbosons. This finite number has profound effects,
and leads to numerous predictions that are different from those of the
geometric analogues of IBA structures.
• The states of this boson system result from the distribution of the
fermions in s and d pairs, and thus depend only on the s and d boson
energies and on interactions between bosons. These interactions are
assumed to be simple (at most two-body).
A fundamental feature of the IBA that results from these assumptions is its
group theoretical structure. Since an j boson (J = 0) has only one magnetic
substate and a d boson (J = 2) has five, the s-d boson system can be looked at
mathematically as asix dimensional space. The basis states span that space. It
turns out that such a system can be described in terms of the algebraic group
structure U(6). As we shall discuss at length, such a "parent" group has
various subgroups and different "decompositions" of U(6) into sequences of
subgroups that lead to different symmetries (dynamical symmetries, to be
exact). There are three of these symmetries that are physically interesting,
known by the labels U(5), SU(3), and O(6). Each has specific, characteristic
properties and a definite geometric analogue. Actual examples of nuclei
manifesting these symmetries have been identified.
This symmetry structure is central and critical to the IBA. Many predic-
tions can be obtained analytically by powerful group theoretical (algebraic)
methods, rather than by tedious numerical diagonalizations. Moreover, the
group structure keeps the underlying physical picture close at hand. Even
when analytic results are not obtainable (for nuclei with structures "intermedi-
ate" between two symmetries) the symmetries act as benchmarks or touch-
stones that provide a physical backdrop and a simpler starting point for de-
tailed calculations.
As Broglia and others have pointed out, the twin aspects of the grounding of
the IBA in fermion pairs, and its symmetry structure with geometrical ana-
logues, confers on the model a "Janus-like" character. On the one hand, it
looks to the shell model for its microscopic justification (is it a reasonable trun-
cation to ignore all configurations except those corresponding to J = 0 and J=2
valence fermion pairs?) and for the ultimate derivation of its parameters and
their systematics. On the other, it leads to a picture of the nucleus, and specific
predictions, very closely allied to macroscopic geometrical collective models.
To summarize, the two essential distinguishing features of the IBA are its
symmetry structure resulting from the s-d boson truncation and its emphasis
on the valence space, with explicit recognition of the finite number of valence
nucleons. The first leads to its algebraic formulation and to the dynamical
symmetries so intimately associated with the model, while the latter leads to
200 Collectivity, Phase Transitions, Deformation

many key predictions, often different from otherwise closely related geometri-
cal models. As we shall see, it confers a microscopic aspect on an otherwise
basically phenomenological model.
We now present a simplified outline of some key elements of the IBA-1
model (protons and neutrons treated together). We first discuss the bosons
and the basis states that can be constructed from them, and then a suitable
IBA-1 Hamiltonian. We then turn to a discussion of the group theory of the
IBA and its symmetry structure. Finally, we consider realistic (nonalgebraic)
calculations for actual nuclei and a simplified approach to many of these, the
so-called consistent Q formalism. Throughout, we give a number of concrete
examples of IBA predictions and stress the relationship to the geometrical
models discussed earlier in this chapter.
The basic entities of the IBA are s(J = 0) and d(J = 2) bosons, which are
assigned energies e and ed. (Note that it is conventional in the IBA literature
to use L for angular momentum both for the individual bosons, s and d, and for
the total spin of a state. Here, we keep to the convention of this book and use
J for these quantities.) A given nucleus with N + Nn valence protons and
neutrons (each counted to the nearest closed shell) has N = (Nf + Nn)/2 s and
d bosons. For example, 152Sm has N = 6 + 4 = 10 and both ^Ba^and 196Pt have
N = 3 + 3 = 6. No distinction is made whether the valence nucleons are par-
ticles or holes. Ground and excited states are formed by distributing the bo-
sons in different ways among s and d states and coupling them to different total
J. The level structures that result depend on these distributions and couplings.
The simplest situation is to imagine all TV bosons in s boson states. By
convention, es = 0, the absolute ground state. The lowest excited state will have
(N - 1) s bosons, one d boson, and an energy E = ed. The next states, in this
simplest case, will be a group with two d bosons (nd = 2). Clearly, as in the
phonon model, the two d bosons can couple to J = 0, 2, 4. This triplet will be
degenerate. Higher d boson multiplets will also occur up to nd = N. This is a
purely harmonic spectrum identical to the harmonic vibrator except for the
limitation due to finite boson number.
Since the IBA is configured explicitly in terms of s and d bosons, most of the
formalism is phrased in terms of creation and destruction operators for these
entities, s, s f , d, d f , and combinations thereof. The basic rules for operating
with these are the same as for the phonon operators b, b f used earlier in this
chapter (Eqs. 6.2-6.4). The Hamiltonian for the harmonic system just de-
scribed, is simply

that is, the energies are E = end, where for simplicity here and henceforth we
drop the subscript "d" on e.
Different states in a multiple! can be distinguished by their angular
momentum / and by the number np(n&) of d bosons coupled pairwise (tri-
pletwise) to J = 0. Sometimes one specifies not nft, but the number of bosons
not coupled to / = 0, and denotes this "boson seniority" quantum number by v.
These states form a convenient basis set for the IBA and are illustrated in Fig.
6.24. Note that for nd> 4, more than one state of a given / can occur.
 Collective Excitations in Even-Even Nuclei 201

Having defined the basis states in this way, we can now consider more
general IBA Hamiltonians composed of creation and destruction operators
for s and d bosons limited to a maximum of two-body (boson) interactions. If
we keep only those terms relevant to excitation energies (i.e., if we ignore
terms contributing to binding energies), we can write

where e ', Cp v2 and v0 are six free parameters. As we have discussed, the first
term simply counts the number of d bosons and multiplies it by a d boson
energy. This gives the unperturbed energy of a state with nd noninteracting d
bosons. The second group of three terms introduces interactions between
pairs of d bosons that depend on the angular momentum to which they are
coupled but that do not change the relative numbers of s and d bosons nor mix
the basis states. The other terms have the property of changing the number of
d bosons by An^ = ±1, ±2. These terms mix different basis states of a given /
and, as in the analogous case of the shell model, it is this configuration mixing
that leads to a build-up of collectivity and to the appearance of rotation-like
behavior.
One often sees another equivalent form of the IBA Hamiltonian,

where
202 Collectivity, Phase Transitions, Deformation

The operators in Eqs. 6.51 and 6.52 are convenient combinations of those in
Eq. 6.50 that have simple physical interpretations in terms of, for example,
boson pairing and quadrupole operators. The most important point to note in
Eq. 6.51 is the And character of the various terms; those in nd, J2, T32 and T42
have &nd = 0, P TP has And = 0, ±2 contributions, while Q f has An, = 0, +1, ±2
parts.
An important aspect of IBA predictions focuses on E2 transitions. The
relevant operator, T(E2), is simply related to Q in Eq. 6.52, by

where eg is a boson charge similar to the effective charge for fermions and is
often treated as a free parameter. In the original IBA formalism, the parameter
X in Q is fixed at % - - fJ/2 in the Hamiltonian and treated as a free para-
meter in T(E2). An alternate formalism, the consistent Q formalism (CQF),
uses the same % in both //and T(E2), which leads to certain simplifications and
to a clearer physical picture of this model. This formalism will be discussed
shortly.
We have stated that the s and d boson structure of the IBA leads to a six
dimensional space and hence to a description in terms of the unitary group
U(6). We shall not delve into the group theory of the IBA in any detail, but a
few ideas are useful to understand how the symmetries so characteristic of this
model arise. Much of the following discussion is based on a recent review by
the author and D.D. Warner to which the reader is referred for additional
material on the IBA and its literature.
The basic concept underlying the group theory of the IBA is that of the
"generators" of a group. These are sets of operators that "close on commuta-
tion" (i.e., the commutator of any pair [A, B] = AB - B A either vanishes or is
proportional to another member of the group, or a linear combination thereof).
For the IBA, the 36 operators sfs, s^, d^/s and (d^dp7 where / = 0,1,2, 3,4
and l/il <J satisfy this condition and are the generators of U(6). As an example,
we show this closure for the particular pair dfs and s ts.
Using Eqs. 6.2-6.4 we have

or, since ns is just a number and is factorable

or
 Collective Excitations in Even-Even Nuclei 203

The other commutators can be similarly evaluated and indeed close on com-
mutation. This set of 36 generators of the group of transformations of U(6) is
said to form a Lie algebra.
Another key concept is that of a Casimir operator of a group. This is an
operator that commutes with all of the generators of the group. Such opera-
tors can be composed of linear or higher order combinations of the generators
and are appropriately called linear, quadratic,... Casimir operators.
The linear Casimir operator of U(6), which commutes with all 36 genera-
tors, is the total boson number operator N = d T d+ s*s whose eigenvalue is N.
This result follows trivially from the fact that all 36 combinations of the s and
doperators must conserve the total boson number. For example,

Suppose now that some smaller set of operators also closes on itself under
commutation. This set forms the generators of a smaller subgroup of U(6). It
will have linear and/or quadratic Casimir operators associated with it that
commute with all the generators of the subgroup. There are several subgroups
for U(6), so the reduction process continues until the rotational subgroup O(3)
is reached.
It is now necessary to find the quantum numbers that label the states. In
general, the generators of a group may change some quantum numbers (e.g.,
nd) but there will be one (or more) that are not changed by any of the
generators. For U(6), the 36 generators always conserve N. The set of basis
states that have a particular fixed value of an unchanged quantum number (or
numbers) is called an irreducible representation of the group.
Since the generators of a given group cannot connect different irreducible
representations, the Casimir operator(s) of a group that commute with all the
generators by definition must be diagonal and therefore must conserve all
quantum numbers, including those of the subgroups. Indeed, each Casimir
operator has eigenvalues that are functions only of the conserved quantum
numbers of the particular subgroup. Thus we have the central result that a
Hamiltonian consisting of Casimir operators of a group and subgroups cannot
mix different representations of any of the groups involved. Furthermore, its
eigenvalues are simple linear combinations of its component Casimir operator
eigenvalues and are functions of the quantum numbers characterizing each
group and subgroup. Since the quantum numbers characterizing a subgroup
are constant for all the states of the particular representation it defines, all the
states of that representation must be degenerate. This degeneracy is broken
only by the next step in the chain, which subclassifies the levels according to
another quantum number (for a subsequent subgroup). This whole process is
illustrated for one of the group chains (the so-called O(6) limit) of the IB A in
Fig. 6.25. The precise meanings of the quantum numbers and eigenvalue terms
will be clarified shortly. The key point here is the successive degeneracy
204 Collectivity, Phase Transitions, Deformation

Fig. 6.25. Illustration (using the 0(6) symmetry) of the successive degeneracy breaking in a
dynamical symmetry group chain (Casten, 1988a).

breaking and the classification of sets of states (representations) of each

subgroup by specific quantum numbers. Another important result is now also
obvious: a transition operator consisting of generators of a given group or
subgroup cannot connect states in different representations of that group.
This leads to many essential selection rules.
A central task in developing any group chain or group reduction scheme is
to identify the quantum numbers that label the irreducible representations of
each subgroup. This is the basic procedure followed in the algebraic treatment
of the IBA. Group chains are constructed starting from U(6), where all the
states are degenerate for a given value of N, and ending with O(3). A
Hamiltonian for any such chain is written as a sum over the Casimir operators
of the subgroups of the specific chain, and is therefore diagonal in a basis
defined by the corresponding representation labels. Each step in the chain
reduction introduces one or more free parameters (coefficients of terms in //)
into the eigenvalue expression and requires one or more quantum numbers to
distinguish the representations of the particular subgroup; it also breaks a
previous degeneracy. Thus, the solution of the eigenvalue problem for such a
chain reduces to that of the (known) eigenvalues of each of the Casimir
operators.
The structure defined by such a Hamiltonian is referred to as a dynamical
symmetry. One of the elegant aspects of these symmetries is that the excitation
 Collective Excitations in Even-Even Nuclei 205

energy spectrum can be written down immediately and each state can be
labeled by appropriate quantum numbers even though these symmetries may
correspond to a complex physical situation and, in terms of Eq. 6.51, to a
complex Hamiltonian. Since transition operators can often be written in terms
of the group generators, transition selection rules appear naturally, and the
rates for allowed transitions can be written analytically. Moreover, many
ratios of transition rates depend only on general characteristics of the symme-
try (group chain) and are parameter free. This should not be surprising: the
Alaga rules for E2 branching ratios in deformed nuclei are a familiar geomet-
rical analogue.
Returning now to the basic problem in the IBA, there are exactly three
group chains of U(6) that end in O(3), which is the rotation group. This group
is a necessary subgroup in any physical chain because it provides for rotational
invariance. The three group chains can be written, along with their relevant
quantum numbers (see discussion to follow) as:

We now discuss each of these symmetries in turn.

I. U(5)
The U(5) symmetry is the IBA version of a vibrator. Its representation
labels were already introduced, since this limit provides the basis states used in
most treatments of the general IBA Hamiltonian.
The eigenvalues of U(5) are

where a, /?, 7, and <5are parameters. A harmonic version of U(5) was illustrated
hin Fig. 6.24. Note, however, that U(5) is a very rich symmetry and allows much
anharmonicity. The degenerate multiplets with a given value of nd include
levels with different values of v, J and the energies can depend on these
quantum numbers. The U(5) wave functions, of course, are trivial. Since they
are themselves the normally-used IBA basis states, each wave function has but
a single term. Even highly anharmonic U(5) spectra maintain the same simple
wave functions: the anharmonicity is a diagonal effect on the energies and does
not lead to mixing of the basis states.
An interesting result concerning the anharmonicity follows from the Hamil-
tonian for U(5) written in the form of Eq. 6.50. Here U(5) includes all terms
with And = 0. Although the interactions that break the degeneracies may
206 Collectivity, Phase Transitions, Deformation

appear to be complex, they never involve higher order interactions in d and s

boson operators (i.e., operators such as ((PcPcPXddd)). Thus the anhar-
monicities, whatever they may be, arise from two-body interactions. This is
exactly the ansatz used to discuss the relationship between three-phonon and
two-phonon energies in the anharmonic vibrator model earlier. In neither case
are the results dependent on the details of the interactions, but only on the
number of the interacting entities. Thus, these same relationships, listed in
Table 6.4, characterize the U(5) limit. No choices of U(5) parameters can
violate them.
In each symmetry that we shall discuss, key signatures and tests are pro-
vided by E2 transitions. So it is useful to summarize some typical results for
each.
The operator T(E2) has a term that changes nd by ±1, and a term with
And = 0. If T(E2) is chosen to be a generator of U(5), only the latter term is
used. (Since nd is a good quantum number in U(5), a generator of U(5) cannot
change nd.) The predicted E2 matrix elements would then vanish between
states differing by ±1 d boson, while they would yield nonzero diagonal
contributions (quadrupole moments). This situation is essentially the inverse
of what is expected and observed for vibrational nuclei, making it customary to
use the first term of the E2 operator in the U(5) limit, as this term produces
results similar to those of the geometric vibrational picture.
For example, one obtains the general result

The sum on the left side of Eq. 6.59 accounts for the distribution of strength
from a given initial state if the angular momentum selection rules allow decay
to more than one level of the next lower multiplet. This sum contains more
than one term only for decay of nd> 3 states, and is identical in origin to the sum
in the phonon model expression, Eq. 6.6.
The factor (nd + 1) in Eq. 6.59 is analogous to the phonon model result
proportional to (N H +1). The factor (TV- nd) in the IBA case has no analogue
in the phonon model and arises specifically from the finite boson number. Its
origin can easily be seen. The matrix element (nd, ns s *d I nd+ 1, ns-1) can be
calculated as:

In terms of the number of quadrupole excitations only (i.e., d bosons or

quadrupole phonons) the U(5) limit and the geometrical vibrator are identical.
The 2j+ level has one such excitation, the levels of the 0+, 2+, 4+ triplet have two,
and so on. The difference is that, in such excitations in the IBA, the restriction
to a finite fixed N imposes another complementary constraint on the number
of s bosons, which gives rise to the second factor on the right in Eqs. 6.59 and
 Collective Excitations in Even-Even Nuclei 207

6.60. In the phonon model, a creation (or destruction) of a phonon takes place
in isolation: the transition rate is related to the number of phonons available.
However, in going from one U(5) representation to another in the IBA, the
creation (or destruction) of a d boson must involve the destruction (or crea-
tion) of an s boson to conserve N. As nd grows, there are fewer available s
bosons, and the 5 boson factor ns (or N- n^ decreases. In an (nd+ 1) —> nd
transition in the IBA, larger nd values facilitate the transition (there is more
freedom in choosing a particular d boson to destroy) but the smaller number of
s bosons hinders the transition. These two counterbalancing aspects are
reflected in the two factors in Eqs. 6.59 and 6.60.
Equation 6.59 gives, for the transitions between the lowest levels,

and

The ratio of these two equations gives the useful result

Since U(5) is usually relevant only near closed shells where N is rather small,
differences with the geometric model can be significant. For example, Eq. 6.63
gives R = 1.6 for N = 5, compared to R = 2.0 for the geometric picture. Finally,
when the initial state is the fully aligned J = 2N excitation, the factor (N- nd) is
reduced to unity. This is an example of the well-known cutoff effect in B(E2)
values involving high spin states, which is another characteristic distinction of
the IBA from geometric models.
Whenever some model predicts a symmetry, it is always a critical test to
search for empirical examples. This is particularly true for the IBA since it is
so intimately connected with the concept of dynamical symmetry. Searches for
U(5)-like nuclei naturally focus on those regions where the geometric vibra-
tional model is also appropriate. The nucleus "8Cd has recently been proposed
as a near-harmonic empirical manifestation of U(5). Its level scheme was
shown in Fig. 6.9. The E2 selection rules for U(5) (identical to the vibrator) are
clearly well-satisfied, and the small anharmonicities in the two-phonon states
reasonably account for the three-phonon anharmonicity. It is also possible to
fit many nuclei with more highly anharmonic U(5) energy spectra but strong
disagreements with the U(5) E2 selection rules and branching ratios nearly
always result.

II. SU(3)
This symmetry is the IBA version of a deformed rotor, but with special
characteristics that distinguish it from its geometric analog. The SU(3) limit is
208 Collectivity, Phase Transitions, Deformation

obtained when the Q2 term dominates in Eq. 6.51 (a J2 term may also be
present). Thus

Here, for Q to be a Casimir operator of SU(3), % must equal - V7/2 . We saw

before that Q2 strongly mixes U(5) basis states with And = 0, ±1, ±2. Therefore,
SU(3) wave functions are no longer simple in terms of an expansion in U(5).
On the contrary, they are rather complex, and certainly not very physically
transparent, combinations of many U(5) states. The simplicity of SU(3), or
any IBA symmetry, results from its geometrical structure, from the analytic
nature of many results, and from simple selection rules, despite the fact that
the wave functions, when expressed in the basis of another symmetry, may be
complex. Such complexity signals only that the symmetries are different from
each other, not that one is more complicated than any of the others. Neverthe-
less, one gets an insight into the symmetry structure by explicitly showing some
wave functions in the same basis. This is done for all three symmetries in a
U(5) basis in Table 6.12.
On account of the mixing of basis states with different nd values in SU(3),
the expectation values of the operator (d td)° = nd, are very different in SU(3)
and U(5). They are shown for the ground band states in Fig. 6.26 along with
the values for O(6). In U(5), (nd) = 0,1,2,... etc., up to (n) = N. In SU(3), (n)
is already substantial in the ground state. This has three important effects that
we can see without detailed calculation. First, any effects of finite boson
number will be relatively larger in the SU(3) ground state than in U(5).
Second, since in both cases (nd}maK = N, the expectation value of nd must
increase slower with J in SU(3) than in U(5). This will have important
consequences for certain B(E2) values that, we shall see, will increase more

Table 6.12. Wave functions expressed in the U(5) basis for the first three 0+ states in each limit of
the IB A
Basis States (ndngnA)
State* Limit (000) (210) (301) (420) (511) (602) (630)
U(5) 1 0 0 0 0 0 0
0+i 0(6) -.43 -.75 0 -.491 0 0 -.095
SU(3) .134 .463 -.404 .606 -.422 -.078 .233

U(5) 0 1 0 0 0 0 0
0+2 0(6) .685 .079 0 -.673 0 0 -.269
SU(3) .385 .600 -.204 -.175 .456 .146 -.437

U(5) 0 0 1 0 0 0 0
0+3 0(6) 0 0 -.866 0 -.463 0 0
SU(3) -.524 -.181 -.554 .030 -.114 -.068 -.606
*The states are ordered for pedagogical clarity and not necessarily in the order of increasing energy: indeed, the
T= 3 0+ state in 0(6) (here labeled 0+3) is usually the 0+2 state.
 Collective Excitations in Even-Even Nuclei 209

Fig. 6.26. Expectation values of n in the ground band in the three limits of the IBA (Casten,
1988a).

slowly with J in SU(3) than in U(5). Third, in U(5), (n^ for a given state is
independent of N, while in SU(3), it is roughly proportional to N. This has
enormously important effects on collective E2 transitions.
The SU(3) energies are given in terms of quantum numbers of the group
chain II of Eq. 6.56

where av a2 are the coefficients of the multipole form of the Hamiltonian of

Eq. 6.51. Each set of (A, //) values defines a representation of the subgroup of
SU(3) and corresponds to a set of one or more rotational bands. Each band is
characterized by a quantum number, sometimes denoted K', which is almost
identical to the usual K projection quantum of geometrical shape models.
(Technically, some SU(3) states contain small admixtures of other K values;
this has notable effects on certain B(E2) values since it is a bandmixing effect,
but as far as the wave functions are concerned, it is an excellent (and useful)
210 Collectivity, Phase Transitions, Deformation

approximation to ignore these mixtures and use the usual notation K.) The
rule that determines the K values that occur in a given (A, //) representation is
K = 0, 2,. . . min(A, ju), K even. For typical values of a2 < 0, the ground state band
has (A, n) = (2N, 0). The next representations are (A, fj) = (2N - 4, 2) with
K = 0, 2 bands, (2N - 8, 4) with K=0, 2, 4 bands; and (2N - 6, 0) with a single
/£ = 0 band. These states are illustrated in Fig. 6.27. The similarities to a
deformed rotor are clear: we see sequences of states resembling a ground state
band, /3 and y vibrational bands, and bands that can be characterized as the
Pfi(K = 0), j8y (K - 2), and 77 (K = 0, 4) two-phonon intrinsic excitations.
However, we note two specific features, exemplified by the (3 and 7 bands, that
distinguish SU(3) from a general deformed rotor and act as characteristic
signatures for the symmetry. They are schematically shown in Fig. 6.28. Since
the P and 7 bands appear in the same representation, states of the same spin of
these two bands must have the same energies. Thus, SU(3) is a special case of
a deformed rotor with degenerate /3 and 7 bands. We stated that a transition
operator consisting of the Casimir operators of a subgroup cannot connect
different representations. Therefore, in the SU(3) limit, the E2 operator with
X -- VT/Z cannot lead to transitions from either the /3 or 7 bands to the ground
band! This is in direct contrast to the usual picture of harmonic collective ft
and 7vibrations in deformed nuclei. Moreover, since these bands are in the
same representation, the E2 operator leads to allowed, collective 7-* j3E2
transitions, again violating the AN h = ±1 selection rule of geometrical models.

Fig. 6.27. Characteristic spectrum of SU(3) (Casten, 1988a).

 Collective Excitations in Even-Even Nuclei 211

Fig. 6.28. Characteristic signatures of SU(3) (Casten, 1985b).

Given decades of success of the latter model, this would appear to be an

argument against the IBA. However, neither the issue nor the data is as simple
as at first appears, and both approaches can be made more or less mutually
consistent. We will discuss this shortly to help us to better understand the IBA.
There are three more distinguishing characteristics of the SU(3) symmetry
shown in Fig. 6.28. Since there is no connection between representations,
there can be no y-g or /3-g bandmixing. Zr(Zp) are effectively zero in SU(3).
And, although both /J -» g and y-> g E2 transitions vanish in SU(3), they have
a finite ratio. Specifically,

Finally, we see from Eq. 6.65 that if we substitute (A, n) = (2N, 0) and
(2N-4, 2) for the ground and y intrinsic excitations, we obtain the energy
difference EJJ) - E (J) <* (2N-1)—the yvibrational energy increases with N
in SU(3) towards midshell.
We note that in Fig. 6.28, the collectivity of /3 -> /transitions is enclosed in
brackets because it persists even with large SU(3) symmetry breaking and, as
such, cannot properly serve as a specific signature of the limiting symmetry. It
does, however, distinguish the IBA from harmonic geometrical models.
A few specific results of SU(3) are useful to cite for practical applications.
The parameters a2 and a, of the eigenvalue expression may be written in terms
of specific level energies by inserting appropriate values of K and //. One
obtains, for example,
212 Collectivity, Phase Transitions, Deformation

Ground band B(E2) values are given by

Hence,

Note that these go as N2 for large N, in direct contrast with the linearity in N
characteristic of U(5). The reason is obvious and has already been hinted at.
The U(5) B(E2) values scale as N because of the (N - nd) factor in Eq. 6.59.
The nd factor is independent of N because a given pair of U(5) states (e.g., 1*
and Oj+) always have the same pair of nd values (e.g., 1 and 0, respectively)
regardless of N. In SU(3), the d f s and s f d operators in T(E2) give factors
involving N from both operators in each pair, because, as we just saw, both nd
and ns increase with N. So, a dependence on N enters twice, leading to the ~Ni
dependence.
Finally, we note, from direct substitution in Eq. 6.67, an interesting result
that we can illustrate by the ratio

The first factor is the rotational model Alaga. rule. The second factor is
(another example of) an N-dependent finite boson number effect, which
means that even in the strict SU(3) limit, B(E2) ratios deviate from the Alaga
rules. Note that the second factor goes to unity as N -> °°. That its predictions
go over into those of the usual geometrical models for large Wis a characteris-
tic feature of the IBA. Many of the unique aspects of the model (such as
allowed, collective /3 -» y E2 transitions) stem directly from the explicit
incorporation of finite N, which in turn, reflects the model's emphasis on the
valence space.
Since SU(3) is such a specific type of deformed rotor, we already recognize
that it does not characterize most deformed nuclei since such nuclei exhibit
nondegenerate ft and ybands, collective /? -»g and (especially) y-> g transi-
tions, and finite Z . (See Fig. 6.29.) Moreover, in the first half of the deformed
rare earth region E (J)-E (/) actually decreases rather than displaying a pro-
portionality to (2N-1). We will consider shortly how the IBA can treat such
nuclei. First we ask if there are any nuclei that do display the limiting
characteristics of SU(3). The answer is (probably) yes, the rare earth isotopes
of Yb and Hf near neutron number N = 104. The empirical evidence is
displayed in Figs. 6.29 and 6.30, where each of the signatures of SU(3) is
approached in the same general N, Z region (and in no other). At the same
time, it is clear that no single nucleus in the N = 104 region displays all the
Fig. 6.29. Empirical evidence relative to four of the SU(3) signatures near N = 104 (Casten,
1985b).

213
214 Collectivity, Phase Transitions, Deformation

Fig. 6.30. Comparison of empirical and SU(3) values for £ 2 + -£2 + 'n l^e rare earl
h region
(Casten, 1988a). N is ttThe boson number. r

SU(3) characteristics. Moreover, some of these same empirical features (high

intrinsic K = 2 energies, weak B(E2: y-» g) values, small bandmixing) also
characterize high-lying noncollective two-quasi-particle excitations. Thus the
evidence is ambiguous, although the author feels that it points toward an
underlying SU(3) character that may be mixed with noncollective degrees of
freedom; the coincidence of signatures is too significant to be dismissed as
fortuitous.

III. 0(6)
The O(6) symmetry is the least familiar geometrically, although it is now
recognized as corresponding to a deformed, axially-asymmetric but y soft
rotor, the Wilets-Jean model. The O(6) Hamiltonian is

and the eigenvalue equation is

 Collective Excitations in Even-Even Nuclei 215

where A = aJ4, B = a/2 and C = a^ - aJlQ (one sometimes encounters a

notation with coefficients A/4 and fi/6). The characteristic quantum numbers
are a for the O(6) group and i for the U(5) subgroup. A typical O(6)
spectrum was used to illustrate the idea of a group chain in Fig. 6.25, and is
shown more completely (for N=6) in Fig. 6.31. The lowest levels (for A, B > 0)
have cr= N, and T = 0,1,2,.... For each value of i 2, there is a multiple! of
states whose degeneracy is broken by the / (/ + 1) Tterm. For T = 2, there are
only 2+ and 4+ levels, and no triplet as in U(5). Major families of O(6) levels are
grouped and characterized according to a, and within each family, by T (and /,
of course). Note the characteristic behavior within a a family: energies of
states with the same T are monotonic in J (usually decreasing with / since
C > 0) and splittings increase rapidly with T. The ground band or yrast levels
increase as T (T+ 3) = (//2)(7/2 + 3) <* /(/ + 6) as in the Wilets-Jean model. As
noted for the corresponding geometric models, this increase is faster than in
U(5) where E(/) <* /, and slower than in SU(3) where E(/)« /(/ + 1). These
relative ground band energies may be summarized

These expressions are identical to those of the vibrator, Wilets-Jean, and rotor
models shown in Fig 6.21. Despite the apparent differences in ground band
energies for each of the symmetries, it is important to recall that the curves in
Fig. 6.21 are defined by the characteristic quantum number nd, r, and / for
U(5), O(6), and SU(3), respectively. For U(5) and O(6), however, there is also
a separate / (/ +1) term in the Hamiltonian and, depending on the strength of
its coefficient, the actual ground band energies in these two symmetries can be
made to resemble each other or SU(3) itself.
In terms of the U(5) basis states, the nondiagonal term in //0(6) is P f P, which
has And = 0, ±2 matrix elements. Thus the wave functions are mixtures in a
U(5) basis, but are not as complex as in SU(3). Table 6.12 illustrates this,
showing that in O(6) the finite amplitudes always differ from each other by the
addition of a zero coupled pair of d bosons for a given state. This implies that
the number of unpaired d bosons is constant for a given state. It is zero for the
O/ and 02+ states, and 3 for 03\ But note that this quantity is just the boson
seniority v, and in fact, rand v are identical. The use of different notations has
historical origins only and no physical content. Indeed, v or rarise in U(5) and
O(6) because both chains involve the same subgroup O(5). This has been the
source of some confusion since the common occurrence of this subgroup
means that many predictions of the two symmetries are identical. Differences
between them do exist, but reside principally in transition rates, which depend
on the detailed d boson structure and occur among higher-lying states belong-
Fig. 6.31. Typical 0(6) spectrum for N - 6.
 Collective Excitations in Even-Even Nuclei 217

ing to higher nd multiplets in U(5) and toa<Nrepresentations in O(6). (We

shall return to this point momentarily).
The fact that P f P has only And = 0, ±2 terms suggests that one might be able
to construct O(6) wave functions in another way. We pointed out that the Q2
term has And = 0, ±1, ±2 components, but that if % - 0, it has only And = 0, ±2
terms. Therefore, the Hamiltonian

also produces O(6) wave functions and spectra. This Hamiltonian has only
two terms, so it cannot give three independent terms in the eigenvalue equa-
tion. This way of producing O(6) thus leads to a special case of Eq. 6.71 where
A - B. Although this is only one of an infinite set of possible A:B relations, it
turns out to be the one empirically observed in O(6) nuclei, suggesting the use-
fulness of this alternate form for //0(6). We will see that this alternative is in fact
one limiting case of the CQF that offers a simplified approach to many IB A
problems. First, however, we need to delineate a few additional O(6) predic-
tions. The E2 transition selection rules are clear from the form of the E2
operator that is a generator of O(6), namely

The allowed transitions must satisfy Acr= 0 and AT = ±1. The first rule is a
direct consequence of the fact that a generator of a group cannot connect
different representations. It means that there are no allowed transitions from
one a family to another. Eventually, excited levels with a< N must decay, but
only by violations of the strict symmetry. This selection rule provides the most
telling contrast with U(5). (See the following.) The AT = ±1 rule is similar to
the AN h or = And = ±1 rules for the geometrical vibrator and U(5) limits.
Naturally, B(E2) values between yrast states are allowed, and given by

from which we get,

h
As in SU(3), B(E2) values in O(6) scale approximately as N2 for large N. The
yrast B(E2) values are illustrated in Fig. 6.32.
There has recently been much discussion of the differences and similarities
between O(6) and U(5). The level schemes in Figs. 6.24 and 6.31 appear very
different. Specifically, U(5) has the well-known two-phonon triplet, while
O(6) lacks the 0+ state. On account of this difference, the first excited 0+ state
in U(5) decays to the 2,+ level, while the first excited 0+ state in O(6) has T = 3
and therefore decays to the 22+ state. These distinctions have been used as
evidence both for or against each of these symmetries. This issue is more
subtle, however. Each figure embodies a specific choice of parameters (a, j8,y,
218 Collectivity, Phase Transitions, Deformation

Fig. 6.32. Dependence of ground band B(E2) values on J in the three limits of the IB A.

8 for U(5) and A, B, C for O(6)); indeed, the U(5) scheme is the harmonic
limit. Actually, both symmetries permit a rich variety of level scheme configu-
rations by appropriate choices of parameters. Since both group chains contain
O(5) and O(3) subgroups, the only real structural differences center on the
O(6) and U(5) parent groups. Thus, it turns out that the energies of the entire
lowest representation of O(6), with a- N, can be exactly replicated in U(5).
To do so, the 02+ (two-phonon or nd - 2) U(5) level must be forced up in energy
above the nd = 3 03+ state in order to reproduce the O(6) decay pattern in which
the first excited 0+ state decays to the 22+ level. Such an O(6)-like U(5)
spectrum would be enormously anharmonic but still valid within the context of
this dynamical symmetry.
The real difference between O(6) and U(5) occurs in two other realms:
absolute transition rates and higher-lying levels. Table 6.12 shows that, though
their energies may be identical, the U(5) and O(6) wave functions are com-
pletely different. Thus, B(E2) values will be different; in particular, the
different expectation values of nd imply different finite N effects. We have
seen an example of this in the yrast B (E2) values in Fig. 6.32. In U(5), the yrast
band experiences greater changes in (nd) from state to state (see Fig. 6.26), and
hence the B(E2) values change more rapidly with spin.
High-lying levels in U(5) have high nd and decay to («/-!) levels. There are
many such (nd-V) states and thus many allowed E2 transitions. In 0(6), high-
lying levels often belong to low T states of a < N representations. These
typically have only one allowed deexcitation transition—that permitted by the
AT = ±1 selection rule.
O(6) nuclei are now well known in two regions: the Pt isoptopes, especially
Fig. 6.33. Comparison of empirical energies and E2 branching ratios in 196Pt with the 0(6) limit. The lower numbers on the transition arrows are the predictions,
the upper are the measured values (Cizewski, 1978).
220 Collectivity, Phase Transitions, Deformation

1%
Pt and the Xe-Ba nuclei near A = 130. A comparison of the 196Pt level
scheme with the O(6) limit is shown in Fig. 6.33. The agreement with all the E2
selection rules is impressive. All allowed transitions are observed and domi-
nate the decays of their respective levels; all forbidden transitions are either
weak or unobserved. At the same time, there are at least three important
discrepancies in energies: the 1= 3 0+ state is not below the 3^ level as it should
be, the splitting among the high i states (e.g., the 0+- 2+- 2+, i = 3 - 4 - 5 states)
is much less than predicted, and the energies in the 7 band are less staggered
than predicted.
The data for Xe and Ba is comparable to Pt: only a= N and TV- 2 levels are
known, but many 1 = 4,5,6 states have been assigned and the O(6) character
extends over a large range of nuclei. There are some striking analogies
between Pt and Xe-Ba. Besides their common manifestation of O(6)-like
characteristic, fits of the O(6) eigenvalue equation in each region show nearly
identical ratios A/B ~ 0.9. Interestingly, this common value is very close to the
special case A/B = 1, which corresponds to the CQF Hamiltonian of Eq. 6.73.
Moreover, exactly the same discrepancies with O(6) that occur in 196Pt are
repeated in Xe-Ba. One of these, the weaker 7 band energy staggering,
provides a useful clue to the nature of the responsible symmetry breaking.
Recall from our earlier discussion of asymmetric rotor models that the rigid
triaxial rotor of Davydov is characterized by staggering exactly opposite in
phase to that of the Wilets-Jean 7 unstable model. The O(6) limit corresponds
to the latter, that is, to a completely 7 independent potential and to a nucleus
whose shape fluctuates uniformly over a range of 7 values from 7 = 0 ° - 60°
such that 7ave = 0° but 7ms = 30°. The staggering data suggests that a realistic
potential for Pt and Xe-Ba might contain some small 7 dependence that
would shift the characteristic 7 band energy staggering pattern slightly toward
the opposite couplings of the Davydov model. It turns out that the introduc-
tion of a very small (3-4%) 7 dependence in the potential with a minimum at
30° corrects not only the 7band energy staggering, but the other discrepancies
as well. This idea is in line with our earlier discussion in the context of Fig. 6.22
in which we suggested that the E ^ / E z f d a t a could be accounted for by a
potential that was predominantly 7 independent but with a small minimum
near 7= 30°.
We commented earlier that a strict distinction between O(6) and U(5) can
not rely on the energies or branching ratios of low-lying levels alone, but
requires absolute B(E2) values or branching ratios for a< N states. Both of
these types of data are abundant for Pt and for Xe-Ba. As an example, Fig.
6.34 shows the decay of the most telling a = (N - 2) = 4 level for 196Pt and
compares this with 6(6) and U(5) predictions. The high nd U(5) levels have
multiple allowed decay routes while the O(6) a<N levels often have only one
such route. The data clearly support the O(6) interpretation.
As we have suggested in each of the preceding discussions, it appears that
good examples of all three IBA symmetries exist. This provides support for
the general structure of the IBA model and, moreover, these limiting cases
 Collective Excitations in Even-Even Nuclei 221

Fig. 6.34. Comparison of the decay of the 1604 keV level in 196Pt with O(6) and U(5) symmetries
(Cizewski, 1978).

offer convenient benchmarks to assess the structure of nuclei similar in charac-

ter to the symmetries but exhibiting some degree of symmetry breaking.
Of course, there is no a priori reason why a particular nucleus should satisfy
a given symmetry or, indeed, why any nuclei should satisfy the particular

Fig. 6.35. Symmetry triangle of the IB A with the coefficients giving each dynamical symmetry.
Fig. 6.36. Empirical (upper numbers) and calculated relative B(E2) values for the Os 0(6) -> rotor transitional region (Casten, 1978).
 Collective Excitations in Even-Even Nuclei 223

Fig. 6.37. Empirical and calculated properties of the transitional Sm nuclei (Scholten, 1978).

constraints of one of these symmetries. Most nuclei do not, yet the IBA is
appealing precisely because it provides a simple way to treat such nuclei. Such
treatments also reveal new aspects of collective behavior not heretofore en-
countered or expected.
To study some of these ideas, we now turn to transitional nuclei between
IBA symmetries. A convenient way to picture the symmetry structure of the
IBA is in terms of the symmetry triangle shown in Fig. 6.35. The three
symmetries mark the vertices. In terms of the Hamiltonian of Eq. 6.51, the
characteristic wave functions of each symmetry are generated by a specific
term whose coefficient is labeled at the appropriate vertex: e for U(5), a2 for
SU(3), and «0 for O(6). We have also seen that for Q to be a generator of SU(3)
one must set# = -V7/2, and that if % = 0, the same Hamiltonian produces O(6)
wave functions. These % labels are also included in Fig. 6.51.
Treating transitional nuclei is trivially simple. Structural evolution from
one limiting symmetry to another merely involves changing the ratio of the
two parameters associated with the two vertices from 0 to °°. For example, a
U(5) -> SU(3), or vibrator -> rotor, transition, such as occurs in the Nd, Sm,
and Gd nuclei near A = 150, is effectuated by varying ela2 and an O(6) -> rotor
transition, as in the Os isotopes, by changing ajar Examples of such structural
evolution are shown for these two cases in Figs. 6.36 and 6.37. The extremely
complex changes in level schemes and transition rates in these phase transi-
tions are rather well reproduced in terms of a variation of a single parameter.
While the SU(3) symmetry represents a deformed nucleus, few deformed
nuclei satisfy its strict rules. Therefore, to properly calculate the bulk of
deformed rare earth and actinide nuclei, one needs to break the SU(3) symme-
try either towards U(5) or O(6). Most deformed nuclei have E^ > E^ We recall
that a high-lying 02+ state is characteristic of O(6). This suggests that devia-
224 Collectivity, Phase Transitions, Deformation

tions from SU(3) in that direction are appropriate. It is indeed possible to

obtain excellent fits to most deformed rare earth nuclei by using finite ratios of
aja2 near - 4. We shall not show such results, but rather turn to an alternate
method, the consistent Q formalism (CQF), an even simpler way of dealing
with these nuclei.
In early applications of the IB A, it was customary to keep % m the quad-
rupole operator in the Hamiltonian fixed at the SU(3) value of-1/7/2, while
treating % in the E2 operator as a free parameter. This is certainly permissible,
even though one might feel slightly uncomfortable using different forms of the
quadrupole operator since it is same quadrupole force that produces deformed
nuclei and collective E2 transitions. This discomfort initiated interest in
pursuing an approach in which consistent forms of Q are used throughout.
This approach, the CQF, is widely used, perhaps due more to its simplicity than
to any philosophical preference. One additional advantage of the CQF is that
it involves fewer free parameters, which makes it easier to establish a relation-
ship between the IBA and corresponding geometrical analogues.
In the CQF, the O(6) -^ SU(3) transition leg in the symmetry triangle is
accomplished by varying % from 0 to - 1/7/2 in both H and T(E2). Now, H has
the form shown in Eq. 6.73

Since J2 is diagonal and has the same effect on any states of the same spin, it
plays no structural role. Thus, if we write

we see that a2 is just a scale factor on energies and has no structural influence.
Thus the wave functions, relative B(E2) values, and relative energies of states
of the same spin are determined solely by x and N. Whereas in the traditional
form of//in Eq. 6.51 the structure and relative energies in an O(6) -> SU(3)
region depend on a0, a2, ay and E2 transitions depend separately on x, now a
single parameter determines all. The only loss of generality is that the O(6)
symmetry approached and obtained in this way is a special case, with A/B = 1.
However, this does not seem to be a deficiency, since it is this special case that
is experimentally observed.
Since predictions of branching ratios and of relative energies (of states of
the same J) depend only on x and N in the CQF, it is possible construct
universal plots for a given N, or universal contour plots against x and N. Some
examples are shown in Figs. 6.38 and 6.39. (In the energy ratio plotted, £2$ is
always subtracted to remove the structurally inconsequential effects of the J2
term.) Figure 6.38 (top) shows the behavior of the /J vibration relative to the 7
band energy. These excitations belong to the representation (N - 4, 2) in
SU(3). As 6(6) is approached, the /J vibration increases rapidly relative to the
7 band. The reason is that the 7 band goes over into a quasi-7 band sequence
starting with the T = 2 2 + level of the a = N family, while the j3 bandhead
Fig. 6.38. Three contour plots in the CQF formalism of the IB A (Warner, 1983).

225
226 Collectivity, Phase Transitions, Deformation

Fig. 6.39. Another CQF contour plot (Warner, 1983).

becomes the yrast 0+ level of the higher-lying a= N - 2 family. Figure 6.38 also
shows that for essentially all ^values, and especially for those values typical of
deformed nuclei (cross hatched box), the /? band lies well above the 7 band.
This is in excellent agreement with the data, but not unexpected given our
discussion. What is more impressive is that the B(E2) ratio in the middle panel
shows that the CQF automatically predicts that B(E2:j8 -» g) « B(E2:y^> g).
In Fig. 2.18 we saw that this feature was one of the most characteristic
empirical properties of deformed nuclei. It is particularly interesting that the
IBA predictions for realistic deformed nuclei come out this way since both
P -» g and y-> g transitions are forbidden in SU(3). However, in the transition
towards O(6), the 7 band levels go over into the quasi-7band (in Davydov
language, into the anomalous levels of the ground band) with collective
transitions to the latter, while the fi band goes over into the low-lying states of
the a = N - 2 family and the ACT = 0 selection rule forbids their decay to the
ground band. Thus they remain weak throughout the transition leg while the
y-> g transitions become collective.
The lowest contour plot in Fig. 6.38 shows one of the most remarkable and
surprising predictions of the IBA, that of collective /? o 7 transitions that are
comparable in strength to 7—* g transitions. This is completely contrary to
normal expectations and to the traditional perception of the experimental
situation. When this prediction of the IBA was initially realized, it was thought
to contradict a wide body of empirical evidence showing systematically
collective 7 —> g and weakly collective ft —> g transitions but no /? <-> 7
transitions. However, given the normal closeness in energy of E? and E and
the fact that E2 transition rates T(E2) <* B(E2) Ef, such transitions would be
extremely weak in intensity even if the matrix elements are large. Recently,
though, extremely sensitive experiments, mostly utilizing the (n, 7) reaction
and powerful 7-ray spectometers installed at an Institute Laue-Langevin in
 Collective Excitations in Even-Even Nuclei 227

Grenoble (see Chapter 10), have shown that collective j3 ^> 7 transitions do
indeed exist and appear whenever the appropriate experiments have been
carried out.
It is evident in Fig. 6.38 that the y-» gift <-> 7 matrix elements generally
increase with N, and in the large N limit, ft <-» 7 matrix elements become
negligible relative to either 7-* g or g —> g intraband transitions, thus recover-
ing the geometrical picture. Of course, geometrical models can be modified
(perturbed) so as to produce ft <-» 7 matrix elements of collective strength
simply by introducing mixing between f} and 7bands. However, as we pointed
out in the bandmixing discussion earlier in this chapter, the B(E2) values so
produced result not as a correction to an unperturbed value (since the latter
vanishes), but solely from the mixing. Thus, all ft -> 7 B(E2) values are
proportional to the same mixing parameter (usually called Z i n analogy with
Z or zp and, hence, their branching ratios must be independent of the amount
of mixing; /J-7mixing provides no flexibility in the branching ratios predicted
in this way and generally disagrees with the data.
The last contour plot, that of Fig. 6.39, is also interesting. It shows a
branching ratio that vanishes in both SU(3) and O(6) limits but is finite in
between. There is no possible path between these limits that bypasses finite
values. Thus, in a totally parameter-free manner, the IBA automatically
predicts that such a branching ratio (and, indeed, many others) will peak in
transitional regions.

Fig. 6.40. Comparison of the data with predictions of the CQFfor two observables. In each case
a straight-line trajectory in % (see dashed line in Fig. 6.39) was used (Warner, 1983).
228 Collectivity, Phase Transitions, Deformation

When carrying out actual calculations, one generally chooses % to repro-

duce a specific energy ratio or branching ratio and then inspects other predic-
tions. However, a more generic viewpoint is obtained by taking the simplest
possible trajectory in % (shown as the straight dashed line in Fig. 6.39) between
O(6) and rotor nuclei. Figure 6.40 shows predictions for such a trajectory for
the branching ratio of Fig. 6.39 and for a branching ratio involving the same
pair of intrinsic states, and compares them with the data for rare earth nuclei.
One does not expect to get exact agreement in such a simplified approach, but
the general pattern of the predictions is in remarkably good accord with the
data.
It is useful to present a more detailed set of predictions for a typical
deformed nucleus situated between SU(3) and O(6) but closer in structure to
the former: 168Er has become, by virtue of the extensive data available for it, the
standard case. Calculated and empirical energies are compared in Fig. 6.41.
The empirical B(E2:y->g) values were shown earlier in a Mikhailov snapshot
in Fig. 6.14. The dashed line is the IBA prediction in the CQF. Table 6.13 also
presents this comparison. In either format, the agreement is outstanding and
demonstrates the automatic incorporation of bandmixing in the IBA.
An interesting feature of this bandmixing is that it decreases with increasing
N. This can be discerned on the left in Fig. 6.40 from the flattening of the

Fig. 6.41. Observed and calculated energies in 168Hr. The J = 4 + , 5", 6' levels of the K = 4 band
predicted near 1.6 MeV are not observed below 2 MeV (Warner, 1980).
 Collective Excitations in Even-Even Nuclei 229

Table 6.13. Relative B(E2) values from the y band in '"Er

', JfJCf Exp IBA (CQF)

2 0,0 54.0 54
2,0 100 100
4,0 6.8 7.6

3 2,0 2.6 2.6

4,0 1.7 1.8
2,2 100 100

4 2,0 1.6 1.7

4,0 8.1 9.6
6,0 1.1 1.5
2,2 100 100

5 4,0 2.9 3.5

6,0 3.6 4.4
3,2 100 100
4,2 122 95

6 4,0 0.44 0.44

6,0 3.8 4.9
8,0 1.4 1.0
4,2 100 100
5,2 69 57

7 6,0 0.7 1.9

5,2 100 100
6,2 59 36
Based on Warner, 1982.

predictions and their asymptotic approach to the Alaga rule value of 0.7 as N
increases. Thus, the IBA automatically predicts a parabolic behavior for Z
against N, minimizing at midshell, that is nearly identical to the empirical
pattern shown earlier in Fig. 6.15.
Of course, the agreement in energy for the higher bands in Fig. 6.41 is
probably at least partially fortuitous since the empirical excitations seem to
have some two-quasi-particle character. Also, the predicted K = 4 band is not
observed below 2 MeV. (It may be necessary to include g bosons in the IBA to
account for this discrepancy.)
From all these results we hope that the power of the IBA as a simple yet
general phenomological model becomes apparent. In other approaches to
collective nuclear structure, one must invoke different models (vibrator, rotor,
etc.), in an ad hoc way to accommodate different structures. In the IBA, a
single framework, embodied in the simple Hamiltonian of Eq. 6.51, encom-
passes all three limiting symmetries, and most intermediate situations as well.
It does so simply through appropriate relative magnitudes of the coefficients e,
av a0, or %, and by diagonalization of a small set of basis states. The "ad-
230 Collectivity, Phase Transitions, Deformation

hocness" is still there, of course, now appearing in the choice of parameters,

but the simplicity of the framework greatly facilitates calculations and helps us
to understand transitional cases in a unified context.
Having dealt at some length with the IB A predictions in the CQF, it is useful
to comment on the physical significance of %. Since the CQF has only one
significant parameter, it is possible to compare IB A predictions as a function of
X with geometrical model predictions as a function of /J or 7. It has been shown
that one can relate /3 and % by the equation

where /?0 is a normalization factor because the scale of fiIBA is undetermined.

Thus we see that, as % varies from - V7/2 —> 0, PIBA goes from fl /J0 —> /J0—that
is, both SU(3) and O(6) are deformed rotors, the latter slightly less so. The j8
- x correlation is relatively minor in importance. The essential structural evo-
lution in the IB A is one of x with 7. This is easy to see by carefully comparing
Figs. 6.19 and 6.42, in which the same three observables are plotted against x
for the IBA and 7 for the Davydov model. (The use of the latter as a point of
comparison is convenient but not valid in the strict sense since the O(6) limit is
7 soft, not rigid; however, the reader will recall our statement that most
predictions of the Davydov model for fixed 7 are nearly identical to those of a
7 soft model with the same 7 ms ) These two figures show that each observable
passes through the same set of values via a similar path. Simply by equating

Fig. 6.42. Three observables, relating to axial asymmetry, as a function of ^ (for N = 16) in the IBA
(compare Fig. 6.19).
 Collective Excitations in Even-Even Nuclei 231

Fig. 6.43. Relation between the geometrical asymmetry y and the IBA parameter % (Casten,
1984). For N - 16 (the dependence on N is weak).

values of each observable, one can assign an effective 7 to each % value. This
7- x correlation is shown in Fig. 6.43. The fact that all three observables show
the same correlation supports the validity of associating each % value with an
asymmetry 7. The picture provided by this correlation is simple: increasing
deviations of % from SU(3) toward O(6) correspond to increasing axial asym-
metry and to larger and larger values of -ymi. Since the IBA never introduces a
minimum in V(j), this increase in ymt can only arise if the potential becomes
increasingly flatter in 7 as % -» 0. Figure 6.44 confirms this by showing the
effective potential V(y) for several % values. This figure shows one other
interesting point. The reader may have noted in Fig. 6.43 that 7*0° for SU(3),
even though this limit is supposed to be that of a symmetric rotor. Figure 6.44
shows the reason. Although the minimum in V(fl occurs at 7= 0°, the potential
is not infinitely steep (for finite N), and zero point motion leads to a finite 7rms.
The body of research relating to the IBA in the last decade is enormous. We
have only summarized a few highlights, emphasizing the symmetries of the
IBA, transition regions, the role of finite boson number, some experimental
tests of the model, and a geometrical understanding of it. We have completely
ignored important topics such as the intrinsic state formalism, which allows
many IBA results to be obtained analytically; the extension of the CQF to the
SU(3) -» U(5) transition region by adding an end to Eq. 6.76; the use of
effective boson numbers, especially in regions where important subshell gaps
occur that might alter the proper counting of valence nucleons; the N Nn par-
ameterization of the IBA; and numerous extensions to the model.
232 Collectivity, Phase Transitions, Deformation

Fig. 6.44. Dependence of the IBA potential on 7 for several % values (Casten, 1984, Ginocchio,
1980).

The two most important of these latter are the IB A-2, in which protons and
neutrons are distinguished and treated separately, and the IBFM, which incor-
porates a single fermion coupled to the boson core so that odd A nuclei can be
calculated. The IBA-2 has its own symmetries and has led to the recognition of
an important new collective mode, the so-called Ml scissors mode, in which
protons and neutrons oscillate (in angle) relative to each other (see Fig. 6.1).
This mode, recently discussed by Richter and collaborators in the rare earth
region, is now known to be widespread near 3 MeV excitation energy in
deformed nuclei. Its existence and properties are closely connected with a new
quantum number that arises in the IBA-2, F-spin, which describes the degree
of proton-neutron symmetry. For many IBA-2 Hamiltonians, F-spin is a good
quantum number. This classification leads to the concept of an F-spin multi-
plet—a group of nuclei, widely dispersed throughout a mass region but dis-
playing similar level schemes. F-spin also facilitates the projection of complex
IBA-2 calculations into simpler, equivalent IBA-1 cases. The development of
boson models for both even and odd mass nuclei raises the possibility of
treating both in a single unified framework. This leads to the concept of bosc-
fermi symmetries or even supersymmetries (often called SUSY's).
These are all extensive topics that merit their own discussion but that arc
well outside the scope of this book. Also beyond that ken are important
 Collective Excitations in Even-Even Nuclei 233

studies extending the basic model to incorporate g bosons. These efforts aim
both to reduce the impact of "boson cutoffs" that lead to reduced B(E2) values
between high-spin states and to account for classes of excitations (e.g., K - 3
and K = 4 bands) that regularly appear among the low-lying intrinsic excita-
tions of deformed nuclei. The incorporation of g bosons has followed two
distinct paths, one in which a single g boson is introduced numerically and
allowed to interact with the s - d bosons, and another in which a full comple-
ment of g bosons is incorporated group theoretically by expanding the parent
group U(6) to U(15).
Other extensions to the IBA involve the inclusion of two-quasi-particle
excitations so that "backbending" phenomena (see Chapter 8) may be treated,
the introduction of higher-order terms such as those cubic ins or d bosons (e.g.,
(d fd f d f ) y (dddy) that incorporate a y dependence in the IBA potential, and
the expansion of the model basis withp and/bosons so that negative parity
states appear.
Finally, a most significant recent facet of the whole arena of algebraic
modelling in nuclear structure is the development of sophisticated approaches
that invoke symmetries in the Fermions directly. Some of these, like the IBA,
emphasize fermion dynamical symmetries in the valence space. Others are
even more general. They use techniques founded in symplectic group theory
to incorporate all oscillator shells simultaneously, thereby affecting important
renormalizations and offering the possibility to describe giant E2 resonances

Fig. 6.4S. An imaginary, but typical, sequence of level schemes and nuclear structure types from a
near closed shell region to a well deformed midshell nucleus.
 234 Collectivity, Phase Transitions, Deformation

and low-lying collectivity without effective charges, in a single coherent alge-

braic framework. Though far beyond the scope of this book to describe, these
more microscopic models are nonetheless active current research areas. For
all this work the reader is referred to the recent literature.

6.7 The Development of Collectivity:

Phenomenology and Microscopic Basis
We have seen that magic nuclei do not display low-lying collective behavior.
Moreover, collectivity and a softness toward deformation go hand in hand as
valence nucleons are added beyond closed shells. In midshell regions of
medium and heavy nuclei, one invariably encounters a large concentration of
deformed nuclei exhibiting rotational behavior and low-lying vibrational exci-
tations of fi, Y, and octupole type.
A schematic view of such a structural evolution, typical in broad brush
strokes of many regions, is shown in Fig. 6.45. The basic trend reflects the
systematics of the Sm nuclei shown in Fig. 6.23. As long as neither type of
nucleon is magic, Ei\ drops and£4 + /£ 2l + ranges from <2 -> 3.33 as nucleons
of either type are added. The structural sequence is shell model -> vibrator -^
transitional —» rotor. In the transitional region, E^\IEi\ usually jumps rapidly
from <2.5 to >3.0. Earlier in this chapter we discussed two kinds of transitions,
one spherical-deformed, the other spherical-y-soft-deformed. £4 \IEi\ values
near -2.5 can occur in either case, but in one case they imply softness to
quadrupole deformation, in the other an extreme softness to axial asymmetry.
Despite the apparent simplicity of Fig. 6.45 and the concepts behind it, real
nuclear systematics is usually much more complex. This is exemplified by Fig.
2.13, which shows the systematics of E*\ in the A = 100 region, and by Fig. 6.46
(left), which shows similar data for the energy ratio Ei:\IE^\ in the A = 150
region. While the general pattern of increased collectivity toward midshell can
be discerned, these plots provide few obvious clues to simple understanding of

Fig.6.46. Normal (left) and N A^ (right) plots of £ 4 +/£ 2 + for the/1 = ISOrcgion (Casten, 1985a).
 Collective Excitations in Even-Even Nuclei 235

the structural behavior. One important clue, however, is provided in Figs.

2.6-2.8, which show the behavior of Ei\ in a set of singly magic nuclei, the Sn
isotopes, and in nuclei with both valence protons and neutrons. We have
commented that the much lower values in the latter strongly suggests that it is
the valence, proton-neutron interaction that leads to softness. Earlier, we
formalized this idea slightly by giving some qualitative arguments why the
T - 0 component of the p-n interaction can induce single nucleon configura-
tion mixing, unlike its T = 1 component, or the like-nucleon interaction, and
why such mixing is tantamount to the development of nonspherical nuclear
shapes. Here, we wish to illustrate and discuss a phenomenological approach
that greatly simplifies the systematics of nuclear transition regions and pro-
vides some insights into the operation of the p-n interaction, and the develop-
ment of collectivity and deformation. (Caveat lector: this is one of the pet
ideas of the author who is, not surprisingly, rather favorably biased toward it.)

The TV,TVn Scheme

The approach is called the TV TVn scheme, and is an attempt to parameterize
nuclear data in such a way as to explicitly emphasize the valence p-n interac-
tion. Suppose we make the assumption that the onset of collectivity, configu-
ration mixing, and deformation in nuclei is simply due to the p-n interaction
and, moreover, that this interaction is fairly long range, orbit independent, and
relevant only for the valence protons and neutrons. Then, its strength will scale
as the product, TV Nn, of the number of valence protons times the number of
valence neutrons. TV p
and TVn are always counted to the nearest closed shells,
J

Fig. 6.47. N Nn plot of E2< for the A = 130 region (compare Fig. 2.14) (Casten, 1985a).
236 Collectivity, Phase Transitions, Deformation

Fig.6.48. Normal (left) and N Nn (right) plots of £ 4 +/£ 2 + for the A = 130 region (Casten, 1985a).

whether these valence nucleons are particles or holes. It is important to

understand that the valence p-n r interaction embodied in the N pN scheme is n

not the total p-n interaction, but the collectivity-and deformation-driving part
of it (primarily the T = Q component). The T- I component must be identical
to the p-p and n-n T = I forces, and as we have emphasized repeatedly, aside
from the pairing interaction, these are repulsive on average and do not lead to
configuration mixing, collectivity, and deformation. (The Sn nuclei in Fig. 2.6
are a classic example of this.)
To illustrate the NpNn scheme, the data shown for the A = 130 region in Fig.
2.14 are replotted, in Fig. 6.47, in terms of N Nn instead of N or Z as is normally
done. Figure 6.48 compares normal and N Nn plots for E^\IE^\ in the same
region.
It is evident that the N Nn plot substantially simplifies the systematics: the
data that fell on several distinct curves before now coalesce so that they can be
described by a single curve for a given mass region. If such behavior is general,
the N Nn scheme offers a phenomenological tool to simplify and unify the
treatment of nuclear systematics, and has a simple underlying microscopic
basis.
However, applying the N Nn scheme to some other regions is less straight-
forward. If there are substantial subshell gaps in the single-particle level
energies, the counting of Np and Nn may be ambiguous. Moreover, the gaps
themselves may also evolve, partly as a consequence of the p-n interaction
itself. If the p-n interaction is expanded in multipoles, the monopole and
quadrupole components generally dominate. The monopole component
P0(cos S) is obviously constant as a function of the angles between the proton
and neutron orbits, and is therefore independent of the total angular
momentum J to which a pair of protons and neutrons are coupled. The
monopole p-n interaction does depend on the relative radial behavior of the
proton and neutron orbits, and therefore on which orbits are tilling in a given
mass region. Its effect is to shift the effective proton and neutron single-
particle energies. As one proceeds through a pair of proton and neutron major
 Collective Excitations in Even-Even Nuclei 237

Fig. 6.49. Illustration of the eradication of a proton subshell gap at Z = 64 as a function of neutron
number due to the monopole p—n interaction.

shells, the specific behavior of the integrated monopole p-n interaction in-
duces distinct patterns of shifts in the single-particle energies that can and do
cause the appearance and disappearance of subshell gaps. Since the radial
integral is largest for orbits with similar shell model quantum numbers (see
Fig. 3.5), the largest effects in heavy nuclei generally occur when spin-orbit
partner orbits, such as g9/2p and g7/2n, are filling. In such a situation, the single-
particle energies of both these orbits are significantly lowered relative to their
neighbors. The consequences can be rather dramatic and are now thought to
be the underlying reason for the very sudden onsets of deformation near
A = 100 and A = 150. The idea is illustrated in Fig. 6.49 for the latter region. As
neutrons begin to fill the h9/2 orbit near N = 90, the strong Ih - In

Fig. 6.50. Contrasting behavior of £2+ for N = 88 and N= 90isotones.

238 Collectivity, Phase Transitions, Deformation

interaction causes the effective single-particle energy of the lhu/2 proton orbit
to decrease and to obliterate the gap at Z = 64. Thus, for N < 90, the effective
proton shell in this region is Z - 50 - 64, while for N > 90, the effective shell is
the normal major shell Z = 50 - 82.
There is some very simple and beautiful empirical evidence for this concept
from the energies of the first 2+ states in the N = 88 and N = 90 isotones. These
data are shown in Fig. 6.50. If the effective proton shell were the normal one
from Z = 50 - 82, one would expect £2; to decrease as the number of protons
were increased past Z = 50 until the near midshell region at Z = 66, after which
£2 f should increase once again. This is precisely what happens for N - 90.
Exactly the opposite behavior, however, characterizes the N - 88 isotones.
(The N = 84, 86 isotones are similar to N = 88.) Without the concept of a
significant subshell gap at Z = 64 for these neutron numbers, such behavior
would be completely incomprehensible. But if one assumes an effective
proton shell Z = 50 - 64 for N = 88, the midshell point is Z = 57, and one would
now expect E^\ to increase for Z between 57 and 64. This is exactly the
behavior observed.

Fig. 6.51. Effective numbers of valence protons and neutrons in the A = 150 transitional region
empirically extracted from g(2 T ] ) factors. The solid lines give the normal dependence of N
against N and the changes in N against N, assuming the Z = 64 gap disappears suddenly between
N = 88 and N = 90 (Wolf, 198?£
 Collective Excitations in Even-Even Nuclei 239

Of course, the idea of an instantaneous disappearance of the Z = 64 gap at

N = 90 is an unrealistically simple scenario. More likely, the dissipation of the
gap is more gradual. Recently it has been shown that nuclear g factors for 2^
states can be used to extract effective valence proton and neutron numbers.
The details of this technique are beyond the scope of this book, but the results
for N ., and N are shown in Fig. 6.51. They support the idea of a rapid but
not instantaneous dissipation of a proton subshell gap as manifested by an
increase in the effective value of N with neutron number for the Ce, Nd, and
Sm nuclei. The solid lines in the figure show the abrupt dissipation scenario at
AT = 90.
A similar story characterizes the A = 100 region. Indeed, it originated in
that region with the pioneering microscopic calculations of Federman and
Pittel; the concept involved is sometimes known as the Federman-Pittel
mechanism. We have seen the effect for this region in Fig. 3.4 as a dramatic
change in single-particle energies between 4091Zr and 50131Sn. The principal
change, a lowering of the Ig^ neutron orbit, is caused by the filling of the
Igg^ protron orbit between Zr and Sn. (Other, like-nucleon effects, arising
from the filling of the neutron 50-82 shell, have a smaller impact since filling a
given major shell is unlikely to alter much the ordering of the same orbits; they
will tend to move in step.) In any case, the lowering of the Ig7/2n orbit lowers
the Ig9/2p orbit in turn, obliterating the shell gap at Z = 38 (or 40) and suddenly
increasing the size of the proton shell from Z=38-50 to Z = 28-50. Nuclei such
as Zr effectively go from N =2ioN =10. Deformation promptly ensues. It
is only because of a mechanism such as this that one can explain the precipitous
drop in £ 2 + for Zr (and Sr) at N = 60 in Fig. 2.13.
This discussion provides a way to link the origin of deformation in terms of
subshell evolution to the structure of intruder states discussed earlier in this
chapter. There, we commented that intruder states appear as low-lying
excitations in some nuclei (e.g., Sn, Pb) because they effectively correspond to
an increase in the number of valence nucleons, and therefore, to an increase in
the p-n interaction. On account of this, they are also more deformed than the
normal states. How low the deformed intruder states go depends on two
factors: the size of the energy gap to be overcome and the strength of the p-n
interaction in the particular intruder orbits involved. Intruder excitations of
the type we considered in the Cd nuclei, or in Pb, correspond to a large gap
(Z = 50, 82) and a moderate interaction. Suppose, however, that the gap is
smaller and/or the interaction stronger. Then it becomes possible for the
intruder energies to drop below those of the "normal" states and become the
(now deformed) new ground state. In this view, the presence of excited
intruders in some regions and the sudden onset of deformation in others are
just two facets of a common mechanism. Such a model, discussed extensively
by Heyde and co-workers in recent years, works and can account at least quali-
tatively for the Cd-Sn and Pb intruders. It also accounts for the spherical-
deformed phase transition near A -100 (Sr, Zr) and A -150 (Ce, Nd, Sm, Gd),
where the Z - 40,64 subshell gaps come into play.
240 Collectivity, Phase Transitions, Deformation

Fig. 6.52. N Nn plot of E2+ f°r the A = 100 region (compare Fig. 2.13) (Casten, 1985a).

The monopole p-n interaction effectively modifies the valence space for the
protons and neutrons. This in turn affects the numbers of protons and
neutrons on which the quadrupole p-n interaction may act. All this suggests
that if one defines Np and Nn based on the known evolution of these subshell
gaps, then NpNn may again be a realistic first-order estimate of the integrated
quadrupole p-n interaction strength. Figure 6.52 demonstrates this for E2{ in
the A = 100 region. Compare Fig. 2.13, which shows the same data in a normal
plot. On the right of Fig. 6.46, the NpNn plot of £4 {/ £ 2 j for the A - 150 region
was included. In these figures the protons shells are taken to be Z = 38-50
(50-64) for N < 60 (N < 90) and Z = 28-50 (50-82) for N > 60 (N > 90). The
effects are dramatic: the extremely complex systematics seen in normal plots
against Nor Z are instantly simplified and the data coalesce into single smooth
curves in NpNn plots. The existence of two curves on the right in Fig. 6.46 is not
surprising; they correspond to the two halves of the proton shell. The circled
points at N - 90, which deviate from the smooth curve, demonstrate the
imperfectness of the assumption of an instantaneous change in proton shell
structure at N = 90. Thus, the N Nn scheme is generalized to all transition
regions in medium and heavy nuclei and provides a simple, yet powerful, guide
to understanding and predicting the systematic behavior of nuclear properties.
It can also be used as a more reliable way of estimating the properties of
unknown nuclei. The reason is that, whereas in traditional plots against N or
Z such nuclei constitute extrapolations beyond the known ranges, the NiNn
values of many unknown nuclei far off stability are actually smaller than those
for known nuclei in the same region. The N Nn scheme converts the normal
 Collective Excitations in Even-Even Nuclei 241

Fig. 6.53. Normal (left) and N Nn (right) plots of g 2+ for the A = 150 region. In the N Nn plot,
only the Z < 64 points are shown. The arrows point to the N Nn values for the extremely neutron
rich nuclei indicated. Predictions for £ 2 + (or other observables) for these nuclei are obtainable
simply by reading off the appropriate curve at these N N values (Casten, 1986).

process of extrapolation into one of interpolation, which is inherently much

more reliable. Figure 6.53 illustrates this, giving as well one more illustration
of the simplification achieved by the N Nn scheme. The vertical arrows point
to the locations of four unknown neutron-rich nuclei in the rare earth region
whose N Nn values are considerably less than those for known nuclei. Since
this plot was constructed, 148Ba has been studied and the predictions verified.
Preliminary reports on another nucleus in this region, 142Xe, also show that the
N Nn scheme provides a much more accurate structural guide.
In the same vein, one can use the N Nn scheme to parameterize collective
model calculations. Normally, in phenomenological models such as geometric
collective models or the IBA, the Hamiltonian contains a number of terms,
ea^h incorporating a free strength parameter. These must be estimated for
each individual nucleus, and therefore, the calculation of extensive sets of
nuclei can involve an enormous proliferation of parameters. If instead, the
parameters are written as functions otN Nn, it is often possible to calculate sets
of 50-100 nuclei with only 4-6 parameters. Moreover, for reasons that are
obvious from the previous discussion, this process automatically give parame-
ter estimates for unknown nuclei further from stability, so their predictions
become interpolative.
Finally, the N Nn scheme has recently been applied to odd A nuclei by
Bucurescu, Zamfir, and colleagues with intriguing results. Energy ratios based
on different initial single-particle configurations probe the p-n interaction for
different orbit combinations. Moreover, N Nn plots in odd nuclei provide a
sensitive signature of the evolution of strongly coupled and decoupled band
structures (see discussion of these in Chapter 8).
Thus far, we have discussed the N Nn scheme for a given region and have
242 Collectivity, Phase Transitions, Deformation

Fig. 6.54. NfNn (left) and P-factor (right) plots summarizing the systematics of £ 4 +/£ 2 + in six
regions of medium and heavy nuclei (based on Casten, 1987).

noted its simplifying power. Further advantages of this approach are evident
when one compares different regions. As we saw in Chapter 2, different
regions appear to behave in entirely different fashion vis a vis the development
of collectivity and deformation: one needs only to glance at Figs. 2.13 and 2.14
or Figs. 6.46 and 6.48 for a dramatic illustration of this. In contrast, N Nn plots
are similar for essentially all regions of medium and heavy mass nuclei. Figure
6.54 (left) collects the smooth curves, drawn through N Nn plots, of the ratio
E^/E2i for six mass regions in heavy nuclei. What appeared earlier to be
radically different behavior from one region to another becomes a set of nearly
parallel curves in N Nn. The similarity in structure of these curves gives
confidence that the N Nn scheme is a meaningful indicator. It also provides
some confidence in extrapolating the curves for a new region once a few nuclei
in that region are studied.
The other feature in Fig. 6.54 is that the curves for different regions are
rather widely displaced: changes in structure are correlated with changes in
NpNn throughout medium and heavy nuclei but the absolute value of N Nn
provides little information. However, a slight modification of the N Nn scheme
puts it on an absolute scale and provides further physical insight. To see this we
consider the parameter

P can be viewed in several ways. It is simply a normalized value of TV yV^. It is

also a measure of the integrated strength of the valence p-n interaction
compared to the valence pairing interaction that scales as the total number of
valence nucleons. Finally, it is the average number of p-n interactions per
valence nucleon. For this reason, P has been called a nucleonic "promiscuity
factor." If we now plot the same data used to obtain Fig. 6.54 (left) against P,
we obtain the results shown on the right of that figure. Individually, the six
curves show exactly the same behavior as their counterparts in the N N^ plot,
but the different regions are coalesced into a narrow envelope, providing a
 Collective Excitations in Even-Even Nuclei 243

unifying framework for understanding the systematic behavior of nearly all

medium and heavy mass nuclei and highlighting, even more than N Nn, the
correlation between the p-n interaction and collectivity.
A slightly more detailed glance at the right-hand side of Fig. 6.54 reveals a
fascinating point directly correlated with our understanding of the residual
pairing and p-n interactions. If we take the value E^ + lEI+ ~3.0 as a measure
of the "transition point" from spherical and vibrational to deformed, we see
that all the regions pass through this value in the narrow range of P values
between 4 and 5. This value, denoted Pcjit, gives a kind of "critical" value that
acts as a signature for deformation. One can formulate a rule: nuclei with
P < Pciit = 4-5 will not be deformed, unlike nuclei with P > Pcrjt. Simple consid-
eration of the formula for P refines this idea: a necessary condition for P > 4 is
that both Nf, Nn > 4, while a sufficient condition for P > 5 is that both
N ,Nn> 10. Thus, no nucleus can be deformed unless there are at least four
valence protons and four valence neutrons and a nucleus must be deformed if
there are at least 10 valence nucleons of each type. Apparent exceptions to this
rule such as the light Hg (Z - 80) nuclei, which have two proton holes relative

Fig. 6.5S. Calculated | S | values for the deformed rare earth nuclei, illustrating the saturation in
quadrupole collectivity. Compare the behavior of empirical B(K2) values in Fig. 5.3 (Casten,
1988b). The legend gives the neutron number corresponding to each symbol.
244 Collectivity, Phase Transitions, Deformation

to Z = 82, are indeed only apparent since the reason deformation sets in these
nuclei is another example of the movement of single-particle energies of one
type of nucleon as a function of the number of the other: as N decreases, the
orbit from above Z = 82 descends across the Z = 82 gap and enters the shell
below. The counting of N should therefore be based on some effective Z
value between 82 and 92.
The value of Pcrit - 4 to 5 is interesting in itself. We know from the energy
gap, 2A, in even-even nuclei that typical like-nucleon pairing interactions
have strengths of V ait ~ 1 MeV. Similarly, p-n interactions are on the order of
200-300 keV (see Chapter 4). Thus, P, which gives the ratio of the number of
p-n interactions to pairing interactions, equals 4-5 at precisely the point at
which the integrated p-n interaction strength begins to dominate the pairing
strength. This provides an appealing physical picture that simply states that
softness to deformation and the phase transition to deformed shapes occur just
when the deformation-driving p-n interaction begins to dominate the spheri-
cal-driving like-nucleon pairing interaction.
The N Nn scheme is based on the rather crude assumption that the p-n
interaction is orbit independent. Since it is not, one might expect situations in
which N Nn is not the best scaling parameter. In fact, we have already seen
evidence for this: most observables are smooth against N Nnbui they are not
linear in N p Nn . The valence *p-n interaction increases with jV N ', but is not
p n

necessarily proportional to it. The most dramatic evidence for this occurs in
the empirical behavior of B(E2:0,+ -> 2t+) values in deformed nuclei. These
increase from the beginning of a shell through the transition region. But,
instead of continuing this increase unabated until midshell, they saturate. This
was illustrated earlier for rare earth nuclei in Figs. 2.16 and 5.3. The reader will
recall from our discussion of the IB A that this B(E2:014 -» 2/) ~ N2 both in
SU(3) and in realistic symmetry breaking calculations for deformed nuclei.
Thus, assuming constant boson effective charges es, such IBA calculations
must disagree significantly with the data near midshell. These and other data
suggest that, although some simplicity will be lost, it might be useful to have a
more refined estimate of p-n strength than N Nn. This can be obtained very
easily by a simple, explicit calculation of the integrated quadrupole p-n
interaction among the valence nucleons. The result, called \Spn\ , is shown in
Fig. 6.55. Clearly, instead otNfNn, one can use these calculated \Spn values to
define effective (N Nn)eff products and therefore effective values of Np and Nn
themselves. If these are then used to recalculate quantities such as
B(E2:Oj+ -* 2j+) values, the observed saturation is excellently reproduced.
If we anticipate our discussion of the Nilsson model in the next chapter, we
can easily explain the behavior of LSJ. (Readers unfamiliar with the Nilsson
model, please forgive this short digression: better yet, return to it after reading
the next chapter.) \S J is obtained simply by integrating (summing) the
products of the individual proton and neutron quadrupole moments for each
Nilsson orbit over all filled orbits up to a given N, Z. Downward sloping lines
in the Nilsson diagram correspond to equatorial orbits, while flat or upward
sloping lines represent more polar orbits. Thus, at the beginning of a shell,
 Collective Excitations in Even-Even Nuclei 245

when both protons and neutrons are filling the equatorial downward sloping
orbits, there will be high overlap and significant contributions to the integrated
quadrupole p-n interaction. Near midshell, however, some neutrons will enter
flat or upward sloping orbits that will have lower overlap with the downward
sloping proton orbits, and vice versa. These, and proton-neutron pairs in
which both particles are in flat Nilsson orbits with zero quadrupole moments,
will contribute little or nothing to further increases in the integrated quad-
rupole p-n interaction strength. This strength should therefore increase line-
arly with TV Nn at the beginning of a shell and then saturate toward midshell.
One last point is relevant before ending this discussion—the concept of a
p-n interaction scaling approximately with NfNn at the beginning of a shell and
then saturating near midshell was originally proposed as a convenient ansatz in
order to provide a simple phenomenological approach to the systematics of
certain collective observables. Very recently, however, Zhang and co-workers
have empirically extracted actual p-n interaction energies for the last proton
and last neutron in a given nucleus and, by carrying out appropriate sums,
obtained the total valence p-n interaction energies for a given nucleus. The
results show exactly the initial linearity with AyV_ and subsequent slower
growth that we have been discussing. That they do not completely saturate
highlights the fact that S n is just the quadrupole interaction. The total p-n
interaction includes the always attractive monopole component whose strength
is monotonic throughout a shell and so provides a continuously increasing
"base" to the total T= 0 p-n strength.
We have seen that the NpNn scheme, and its siblings P and ISJ plots, provide
a simple yet powerful way of correlating a vast amount of systematic data on
the development of collectivity, phase transitions, and deformation in medium
and heavy nuclei. These concepts have an appealing microscopic foundation
in the deformation-driving T~ 0 component of the valence p-n interaction and
its competition with its opposite number, the spherical-driving like-nucleon
pairing interaction. In this way, these ideas, albeit phenomenological, bring
together a number of threads running throughout this book. Through this
phenomenology, we are beginning to develop a unified, coherent view of the
evolution of nuclear structure. This view emphasizes the importance of the
p-n interaction, and its role both in modifying the underlying shell structure
and in inducing correlations, configuration mixing and deformation. What is
needed now is to graduate from phenomenology to a real microscopic theory
of nuclear structure and its evolution that embodies these ideas.
Another key issue is to understand the relation between the obvious
centrality of the p-n interaction and the equally obvious successes of collective
models that make no explicit mention of this interaction. The resolution of this
seeming paradox appears to be twofold. First, the p-n interaction determines
the "mean field," that is, the correlations and deformation of the ground state
(or the base state of a family of states such as the intruders), upon which the
collective excitations are then constructed. Secondly, the p-n interaction,
through its monopole component, affects the detailed distribution and ener-
gies of the underlying single-particle states. As we have seen, this is critical to
246 Collectivity, Phase Transitions, Deformation

the evolution of subshell gaps and therefore to the content of the valence space
on which the quadrupole component acts. Moreover, these single-particle
states are the fodder with which (see Chapter 9) the detailed microscopic
structure (e.g., energies, collectivity) of the vibrational excitations is con-
structed. Thus, the p-n interaction implicitly enters collective models both in
the equilibrium shapes they present and in the single-particle energies used in
obtaining their predictions.
 7
THE D E F O R M E D SHELL M O D E L OR
NILSSON MODEL

The purpose of this chapter is to describe the basic single-particle model

applicable to nearly all deformed nuclei—the Nilsson model. This model is
surely one of the most successful nuclear models ever developed. It accounts
for most of the observed features of single-particle levels in hundreds of
deformed nuclei and is always the first model turned to when new experimen-
tal information on such levels is obtained. It also provides a microscopic basis
for the existence of rotational and vibrational collective motion that is directly
linked to the spherical shell model. It is also very easy to incorporate exten-
sions, refinements, and corrections to it. Essentially a single-particle model,
the Nilsson model has enjoyed particular success in the interpretation of single
nucleon transfer reactions.
Even before discussing this model, we are faced with a conceptual difficulty
arising from the nonspherical shape, or the separation of the motion of an
individual nucleon around the nucleus from rotations of the nucleus itself in
space. These motions can be very different. Imagine a nucleus with prolate
quadrupole distortion and a single nucleon orbiting in an equatorial plane, as
shown by orbit X, in Fig. 7.1. Now imagine that this nucleus can rotate about
an axis perpendicular to the symmetry axis. With rapid rotation about this
axis, the time averaged shape of the core becomes oblate (disc-like). What

Fig. 7.1. Illustration of two single-particle orbits at different inclinations to a prolate deformed
nucleus.

247
248 Collectivity, Phase Transitions, Deformation

shape the orbit of the single nucleon takes then depends on the extent to which
its motion is coupled to that of the core, that is, it depends on the separation of
rotational and single-particle degrees of freedom. A rigorous separation is, in
general, not possible. An approximate separation can be made, however, if the
frequency of the nucleonic motion is much larger than the frequency of the
nuclear rotation, in which case the individual nucleon executes many orbits
during a single nuclear rotation, or, alternately phrased, the nucleus is essen-
tially stationary during a single orbit of that nucleon.
This discussion is not only of formal interest, but alerts us to the possibility
that the separability of these motions may be rather poor for extremely high
rotational velocities. Modern experimental techniques have approached the
limit where the characteristic frequencies for rotational and single nucleon
motion are not distinct. It is then necessary to explicitly incorporate the effects
of the rotation on the single-particle motion. Coriolis mixing is one such effect
that we shall discuss at length, but there are others (relating to the underlying
core shape, the effective single-particle energies, and so on) that are beyond
the scope of this book. Mathematically, the separation of single-particle and
rotational motion greatly simplifies calculations and is the principle reason
why in a body-fixed frame of reference, one evaluates the single nucleon
motion first and later superimposes the rotational motion. (Incidentally, the
same basic type of problem applies to the spherical shell model, in which the
"global" motion is linear motion of the center of mass, which is more easily
distinguished.)

7.1 The Nilsson Model

We will discuss in considerable detail the quantitative properties of the Nilsson
model and in particular the Nilsson energies and wave functions. First,
however, we wish to proceed in a rather unconventional way in order to show
that it is simple to "derive," without detailed calculation, the entire Nilsson
diagram and many of the prominent features of the Nilsson wave functions.
The two necessary ingredients are simply a choice of deformed single-particle
potential, such as a shell model potential with quadrupole deformation /?, and
the recognition that the nuclear force is short range and attractive.
Consider first, then, a valence nucleon in a single / orbit in such a prolate
deformed potential (Fig. 7.1). It will have lower energy if its orbit lies closer to
the rest of the nuclear matter than if it lies at larger distances from it. Clearly,
then, the orbit labeled K^ will be lower in energy than Kr The energy depends
on the orientation with respect to the nuclear symmetry axis. This is contrary
to the spherical shell model where there is no preferred direction in space.
One can specify this orientation by the magnetic substate of the nucleon—that
is, the projection of the total angular momentum on the symmetry axis as
shown in Fig. 7.2. As in our earlier discussion of even-even nuclei, this
quantity is usually denoted by the symbol K. (Technically, Q. is used for the
projection of the single-particle angular momentum on the symmetry axis and
K for the projection of the total angular momentum. However, since the
 The Deformed Shell Model or Nilsson Model 249

Fig. 7.2. Diagram defining the quantities/, K, and 9 in the discussion of the Nilsson model.

rotational angular momentum for low-lying states of axially symmetric nuclei

is perpendicular to the symmetry axis, it contributes nothing to K and therefore
K = Q and is often substituted for it.) The low K values correspond to
equatorial motion near the bulk of the nuclear matter for a prolate quadrupole
distortion, and have lower energy.
One can easily go a step further. Consider the classical orbit angles
corresponding to different A" values. Suppose; = 13/2 (e.g., the im orbit) with
K= 1/2,..., 13/2. Classically, as illustrated in Fig. 7.2, we can approximate the
angle of an orbital plane by 0 = sin-^K/j). These angles are given in Table 7.1.
The interesting feature is that d changes slowly for low K values and rapidly for
high lvalues. Therefore one expects that the difference in energy between
low K values is rather slight and increases rapidly for the higher ones. With
these simple considerations, we can now develop a limited region of a Nilsson
diagram for a single /. This is shown in Fig. 7.3. The characteristic features are
just as we have derived: for /3 > 0, the energy drops rapidly with ft for low K
values and rises rapidly for the higher X values and the separation of adjacent
K values increases sharply with K. We shall see that the energies for small
deformations depend on K2.
The only additional step needed to construct the full Nilsson diagram of
deformed single-particle energies as a function of ft is to combine several;'
values. As discussed earlier, the characteristic feature of a deformed field is
single-nucleon configuration mixing. Therefore, we must now superimpose
this configuration mixing ot different; values on the K splitting just considered.
Table 7.1. Classical orbit angles, relative to the nuclear equator, for/ = 13/2.
K 1/2 3/2 5/2 7/2 9/2 11/2 13/2
0(deg) 4.4 13.3 22.6 32.6 43.8 57.8 90
A6(deg) 8.9 9.3 10.0 11.2 14.0 32.2
250 Collectivity, Phase Transitions, Deformation

Pig. 7.3. Variation of single-particle energies of i^ orbits with different projections K (orientations
6) as a function of deformation (/} > 0, prolate, to the right).

Recalling a fundamental rule of quantum mechanics that no two levels with

the same quantum numbers may cross (an infinitesimal interaction will cause
them to repel when they get sufficiently close) and noting that the only
remaining good quantum number for these orbits is K, it then follows that no
two lines in the Nilsson diagram corresponding to the same K value (and
parity) cross. As two such lines approach each other they must repel (see Fig.
1.9). Thus, it is now possible to incorporate several; values into the Nilsson
diagram and to extend it to realistic deformations where the energies of
different orbits intermingle. This is shown in Fig. 7.4, which gives the Nilsson
diagram for two different regions. Each line, representing a Nilsson state,
starts out straight and is downward or upward sloping according to the angle of
the orbit relative to the main mass of the nucleus. It only starts to curve when
it approaches another level with the same K and parity. The entire structure of
the diagram relies thus on only three factors: K splitting (resulting from the
effects of a short-range nuclear interaction in a deformed field), level-level
repulsion, and the input single-particle shell model energies.
The Nilsson wave functions are equally easy to deduce qualitatively, even
though they are a complex result of a multistate diagonalization. (We ignore
here the phases of the various terms in these wave functions, most of which can
only be obtained by explicit diagonalization.) The interaction that leads to
configuration mixing in Nilsson model is of quadrupole form. (We will discuss
other deformed shapes later.) For very small quadrupole deformations j3, the
nuclear wave functions must be nearly pure in;'; as the deformation increases,
the configuration mixing will increase.
The nondiagonal mixing matrix elements of the quadrupole interaction, not
 The Deformed Shell Model or Nilsson Model 251

surprisingly, tend to mix configurations that differ by two units in angular

momentum and in which the nucleon spin orientation is not changed. For
example, in the 50-82 shell, the dM and s1/2 orbits have large matrix elements
and mix substantially, even though they are slightly further separated than the
d5/2 and dM orbits. The g7/2 and d^ mix more than g7/2 and ds/2do. Likewise, in
the 82-126 shell, the quadrupole matrix element between the pM and f7/2 orbits
is strong. However, the closeness of the energies of the i,a and h9/2 orbits leads
to substantial mixing, even though the matrix element is not favored. There-
fore, combining a regard for the energy separations of different shell model
orbits and the most important quadrupole mixing matrix elements, one can
estimate the Nilsson wave functions, that is, the composition of the wave
functions in terms of amplitudes for different; subshells.
Consider the example of the 82-126 neutron shell shown in Fig. 7.4. As /?
increases, the £,„ and hm orbits begin to mix. We recall that the angle of the
orbital orientation depends primarily on the ratio Klj (8 ~ sin~J KJj ~ Klj for

Fig. 7.4. (a) Nilsson diagram for the Z = 50-82 regions. The abscissa is the deformation parameter
£, which is nearly the same as /3. (Gustafson, 1967).
252 Collectivity, Phase Transitions, Deformation

Fig. 7.4. (b) Nilsson diagram for the N - 82-126 regions. The abscissa is the deformation
parameter e, which is nearly the same as ft. (Gustafson, 1967).

small K). Small angles can occur either because K is low, or for given K,
because; is high. Thus, the energies of the K -1/2,3/2, and 5/2 orbits from the
h9/2 shell decrease in energy faster with deformation than those from the f7/2
orbit. This difference in rate of decrease of the Nilsson energies with deforma-
tion can overcome the small spherical f7/2-h9/2 energy separation. The low
Ktia and h9/2 orbits therefore approach each other, mixing more and more.
However, the two orbits cannot cross and so repel each other, leading to an
inflection point at the value of j8 where they would have crossed. This effect is
very clear for the K = 5/2 and K = 7/2, f^ and h9/2 orbits in Fig. 7.4.
An interesting feature of the Nilsson diagram is apparent if one looks at the
energies past the "pseudo crossing." Starting at large deformations and
tracing back toward /? = 0 the energy of the lowest K = 5/2 orbit is drawn as if
it stems from the f7/2 shell. However, one sees that it actually points directly
back to the h9/2 spherical energy. This reflects the fact that this orbit, for large
deformations, is actually the continuation of the hW2 shell. In effect, while the
 The Deformed Shell Model or Nilsson Model 253

energies do not cross, the wave functions do "exchange" principle wave

function components near the inflection point.
In contrast to these examples, the K = 9/2 orbit from the h9/2 shell is virtually
straight since there is no other nearby (negative parity) / shell with a K = 9/2
component. This reflects the general feature that the Nilsson wave functions
for the highest K values in a given shell are very pure.
The extreme example of this, in the 82-126 shell, is the orbits stemming
from the L,M shell. This shell, with / = 6, has positive parity and lies amidst a
grouping of negative parity orbits. It has been brought down from the next
major shell by the strong spin-orbit interaction in the shell model. Having
opposite parity it cannot mix with any other orbits in the 82-126 shell. There-
fore the wave functions of these unique parity Nilsson orbits are extremely
pure, consisting almost solely of / = 13/2 components even to rather large
deformations. This special structure has many extremely important conse-
quences. The simplicity of their wave functions makes a number of physical
effects particularly simple to understand. Moreover, certain residual interac-
tions such as Coriolis effects are both simple and particularly strong in these
orbits. Finally, many of these same features make the states stemming from
these orbits easily amenable to empirical study.
This discussion of the structure of the Nilsson diagram for a given major
shell can be applied to any shell and the entire Nilsson diagram can be
constructed. The only other point to note is that the higher the shell, the
stronger the effects we have been considering will be, since a particle in a
higher shell is at a larger radius, further outside the spherical nucleus, and
therefore has more to gain energetically, upon deformation, if it is equatorial.
It should be evident by now that it is easy to write down not only the ap-
proximate energies but also estimates of the wave functions of almost any Nils-
son orbit without detailed calculation. We shall give a couple of examples of
this in a moment, but first it is convenient to define the Nilsson quantum num-
bers labeling each orbit. This is also instructive because it highlights the physi-
cal nature of the various orbits. A typical Nilsson orbit is labeled as follows:

The first quantum numbers give the A" value and parity. Inside the brackets
the three quantum numbers are N, the principle quantum number of the major
shell; nz, the number of nodes in the wave function in the z direction (nz is
particularly critical for understanding the structure of the wave function); and
A, the component of the orbital angular momentum along the z, or symmetry,
axis. By definition, K = A + £ = A ± 1/2, where £ is the projection of the intrinsic
nucleon spin on the symmetry axis. Hence, one sometimes sees an alternate
notation [MizAt] or [Nn^A-l], where the arrow replaces K and indicates
whether the spin angular momentum aligns (T) or antialigns (1) with the
oribital angular momentum. The two notations are equivalent, but using K is
somewhat more convenient and common.
For the most common case of a prolate nucleus, equatorial orbits are
nearest the nuclear matter and lie lowest. It is clear from Fig. 7.1 that these
orbits are also those in which the nucleon wave function is most extended in
254 Collectivity, Phase Transitions, Deformation

the z direction. Their wave functions have the largest number of nodes in the
z direction and hence the largest values of n^. We note also that there is a
relation between the permissible values of nz and A such that their sum must be
even if N is even (positive parity) and odd if AT is odd (negative parity).
We are now in a position to label all the Nilsson orbits in a given shell. There
are two ways this can be approached, one by labeling the sequence of orbits
from each spherical; shell and another in which all orbits of a given K value are
labeled according to increasing energy. To make the labeling absolutely clear,
we will illustrate both approaches (for a prolate nucleus). We start from the
fact that the lowest-lying orbit has the highest possible nz. Clearly the maxi-
mum value of nz for a given principle quantum number N is nz = N. The lowest
orbit for a given N has K = 1/2, since this is the most equatorial orbit. Taking
N = 5 for illustration, noting that if K - 1/2, A an only be 0 or 1, and that nz + A
must be odd since N is odd, the Nilsson quantum numbers for the lowest N=5
orbit must be

This Nilsson labeling describes the K = 1/2 orbit stemming from the hn/2; shell.
Although most of the N = 5 orbits occur in the 82-126 major shell, the hllfl orbit
is the one that is pushed down by the spin orbit interaction into the lower, pre-
dominantly N = 4, 50-82 shell. Continuing for the other orbits from the hn/2
shell, the next has K= 3/2. Since its orbital orientation is slightly more inclined
away from the equatorial plane, it is less extended in the z direction and has the
next lower «z value. Its Nilsson quantum numbers are 3/2 "[541]. Again, the A
value is fixed by the requirements that nt + A is odd and A = K ± 1/2. The rest
of the hm-based orbits are then 5/2~[532], 7/2-[523], 9/2-[514], ll/2-[505].
The next N = 5 orbits are in the 82-126 shell proper, and stem from the f7/2
parent (although their actual wave functions contain large h9/2 amplitudes).
The K = 1/2 orbit, having higher energy than the K = 1/2 orbit from the h11/2
shell, must be less extended in the z direction and must have a lower nt. Its
Nilsson quantum numbers are then trivially, 1/2"[541]. The K = 3/2 orbit is
3/2"[532]. One can continue filling the entire shell in this way, and it is easy to
reproduce the labels shown in Fig. 7.4.
An alternative way is to proceed not by energy for a given parent; shell but
by lvalue. For example, the sequence of K = 1/2 orbits, starting with that from
the hn/2 orbit, will be (recall that A = 0 or 1)
l/2-[550], l/2-[541], l/2-[530], l/2-[521], l/2-[510], and l/2-[501]
The unique parity orbit in the 82-126 shell is the lowest N = 6 orbit, and there-
fore its Nilsson quantum numbers must be l/2+[660)], 3/2+[651] 13/2+[606].
Let us now consider the wave functions in more detail. These wave
functions can be written in many forms. Because they involve single-;' configu-
ration mixing, they can be expanded in a spherical basis. For many purposes
the easiest and most physically transparent form is one that the previous
discussion anticipated, an expansion in shell model orbits specified by their;
values. Thus, we write
 The Deformed Shell Model or Nilsson Model 255

where the 0. are solutions to the spherical independent particle model and the
C. are expansion coefficients. Using this language, it is easy to make at least
crude estimates of the actual Nilsson wave functions. For example, for ft = 0.23
(e<= 0.95 /?= 0.22) the 5/2-[523] orbit will have a wave function that is predomi-
nantly h9/2, with the next largest component f7/2. Crudely, we can estimate that
Y (5/2~[523j) = 0.507/2 + 0.809/2 + .... We emphasize that this is at best a guess as
to the Nilsson wave functions and that the phases are arbitrary. We engaged in
this exercise simply because it is useful to have at least some, albeit crude, a
priori feeling for the structure of these wave functions. It will also remove
some of the mystery from the actual Nilsson wave functions. To take another
example, consider that the 3/2~[532] orbit for a deformation of /3 = 0.23 is just
past the inflection point with the 3/2'[521] orbit. Its wave function should be
roughly equal admixtures of f7/2 and h^ components. In contrast, the wave
function for the same orbit for a deformation of P~ 0.05 would be largely f7;2.
The wave functions near the top of the 82-126 shell are particularly simple:
V (1/2~[501]) is dominated by a p1/2 component. Near midshell, the situation is
somewhat more complicated since a given wave function will contain compo-
nents from/shells both above and below it. A particularly nice example is the
1/2~[521] orbit. Careful inspection of the Nilsson diagram shows that although
it starts out from the p3/2 shell, it soon mixes with the K = 1/2 f5/2 orbit,
undergoing a virtual crossing before /?= 0.1. At this point its wave function is
a very thorough mixture of 0M and 05/2 components. Continuing to larger
deformations, at ft ~ 0.2 there is another inflection point due to interaction with
a combination of the K = 1/2 orbits from the f^ and hM shells. Therefore, we
might expect at /? <= 0.25 that iff (l/2-[521j = Cm
with all four C. values substantial in magnitude.
As we anticipated, the unique parity orbits are extremely pure, and increas-
ingly so as K increases. For example, the 13/2+[606] orbit must (assuming no N
mixing) be pure ; = 13/2, while v^(l/2[660]) » 0.95 0m + 0.3 09/2.... (These
numbers are only rough estimates but do embody the physics and reflect the
actual amplitudes from real calculations.) Note that the only other significant
contribution besides i13/2 comes from the g9/2 orbit, which differs by two units of

Table 7.2. Nilsson wave functions (C coefficients) for some N = 5 orbits

/
^N«2A] 1/2 3/2 5/2 7/2 9/2 11/2
3/2-[532] 0.234 0.369 -0.560 -0.651 0.268
5/2~[523] 0.237 -0.472 -0.826 -0.196
7/2-[514] 0.323 0.938 0.128
1/2 [521] -0.510 0.345 0.473 0.431 0.444 0.120
5/2'[512] -0.023 0.836 -0.515 0.157
1/2 [510] 0.021 -0.676 0.586 -0.343 0.277 0.067
3/2'[5l2] 0.379 0.815 0.283 0.327 0.063
7/2-[503] 0.937 -0.336 0.099
9/2-1505] 0.998 0.071
m-[SQi] -0.821 -0.361 -0.411 -0.122 -0.104 -0.019

*5 = 0.22, K = 0.0637, n = 0.42.

256 Collectivity, Phase Transitions, Deformation

angular momentum and has its spin and orbital components aligned in the
same way as the i1M.
Table 7.2 gives examples of some actual Nilsson wave functions for a typical
Nilsson potential (8= 0.22, j3=0.23). Inspection of the table shows that all our
guesses as to the structure are semiquantitatively correct. Of course, fine
details are beyond this discussion and some amplitudes are more difficult to
intuit a priori.
It is worth pausing here to reflect on and to re-emphasize what we have
done. Without any detailed calculation whatsoever, using only simple consid-
erations of the attractive nature of the nuclear force and the nuclear shapes
involved, we have essentially "derived" the entire Nilsson diagram, the Nilsson
energies, and the asymptotic Nilsson quantum numbers. We have also dis-
cussed the basic structure of the wave functions.
At this point, however, we can obtain a deeper understanding of the Nilsson
model and diagram and of the role of the quantum number, n , by a slightly
more formal approach.
To begin, we consider the Nilsson Hamiltonian for a single-particle orbiting
in a deformed potential and inspect two instructive limits, corresponding to
small and large deformation. Actually there are many Nilsson-type Hamil-
tonians incorporating many variants of the single-particle deformed potential.
Various authors have used deformed harmonic oscillator or modified har-
monic oscillator potentials, Wood-Saxon potentials, and others. The differ-
ences reside primarily in details that do not concern us here so we will content
ourselves with the modified harmonic oscillator originally used by Nilsson.
The Nilsson model is a shell model for a deformed nucleus. It provides a
description of single particle motion in a nonspherical potential,
V - V0(r) + V2(r) P2(cos#). The original and basic form incorporated only
quadrupole deformed axially symmetric shapes. An appropriate single-par-
ticle Hamiltonian for a nucleus with symmetry axis z is:

where (ox, (Oy, and a>^ are one-dimensional oscillator frequencies in the x, y, and
z directions. This Hamiltonian satisfies the eigenvalue equation //y/. = £.y/-.,
where y/. is a Nilsson wave function written in the form y.= £.C.'0.. The / 2 and
/ • s terms ensure the proper order and energies of the single-particle levels in
the spherical limit (/3 = 0).
Although the form of the Hamiltonian in Eq. 7.2 is useful, it is also
convenient to introduce an alternate version written directly in terms of a
nuclear deformation parameter 8~ 3/2 ^ 5/4 K p~ 0.95 p. To do this, one writes
the frequencies as
 The Deformed Shell Model or Nilsson Model 257

where co0 is the oscillator frequency (ftfi>0 = 41A~113) in the spherical potential
with 5=0. It is assumed that the nuclear volume remains constant as a function
of t»0. Therefore, one has the condition co^co cot constant or

For positive deformations 8 or /i > 0 (prolate shapes), a>i decreases with

increasing deformation while (£>f and ca increase. This is physically reasonable
since an increasing prolate deformation elongates the nucleus in the z direc-
tion. This increases the "length" of a circumferential route and therefore
lowers the frequency of orbiting in this direction. In contrast, the nucleus is
"squeezed" in the x and y direction so the orbit frequencies can be larger for a
given energy.
Inserting these definitions into Eq. 7.2 allows us to rewrite the Hamiltonian
in terms of the operator r2Y20 as follows

The two equivalent versions of the Nilsson Hamiltonian in Eqs. 7.2 and 7.5
allow us to understand the structure of the model in the limits of large and
small deformations, respectively. Note that, in the literature the 1 • s and I 2
terms are usually expressed in terms of parameters K= C/2ft6)0 and /LL = 2D/C.
K typically takes on values around 0.06 and ju varies from 0 to =0.7.
For small deformation,;' is approximately a good quantum number. Equa-
tion 7.5 consists of a Hamiltonian for an isotropic oscillator with 1 2 and I • s
terms plus a perturbation proportional to ^Y^. The former part gives the
spherical shell model energies and is spherically symmetric. The eigenstates of
this Hamiltonian can be labeled by the quantum numbers Nlj and m of the
spherical single-particle states. Treating the Y20 term as a perturbation, the
shift in energies relative to 8 = 0 is

We can evaluate this by separating the radial and angular parts and using the
relation for a harmonic oscillator potential that

Evaluating the matrix element of the spherical harmonic Y20 gives the final
result for small 8

where we have replaced the projection m with K, the projection of the total
angular momentum on the z axis.
This simple result has three facets that account for the structure of the
Nilsson diagram for small deformations:
258 Collectivity, Phase Transitions, Deformation

• There is a proportionality to 6, the quadrupole deformation.

• The shifts display a dependence on K*.
• They depend linearly on the oscillator quantum number N.
We have seen exactly these features in our intuitive derivation and in the
Nilsson diagram, especially for unique parity orbits for which / is a good
quantum number out to rather large deformations.
Another direct implication of Eq. 7.8 is that for 8 > 0 there are more
downward sloping than upward sloping orbits. For; > 1/2, the [lIP-iQ + 1)]
term is negative, giving downward sloping orbits (since 3/4 -;'(/ + 1) is negative)
if

and upward sloping for K > 0.65;. For example, for ;' = 13/2, orbits with
K = 1/2, 3/2,5/2, 7/2 should be downward sloping and K = 9/2,11/2, and 13/2
upward sloping. This feature is indeed displayed by the exact numerical diag-
onalizations depicted in the Nilsson diagram of Fig. 7.4. Note the interesting
physical correlation here. The angular orientation of an orbit to the symmetry
axis is approximately given by sin0= Klj and K/j-Q.65 corresponds to 0= 40°.
Inclinations greater than these are unfavored energetically by a prolate quad-
rupole deformation.
The dependence on N implies that the slopes of the energy levels in a
Nilsson diagram are steeper for larger N. Thus, heavier nuclei are easier to
deform than lighter ones. We commented implicitly on this N effect earlier
and can now explain its physical origin a bit more precisely. A nucleon in a
high oscillator shell will have a larger average radius [indeed, we just utilized
the fact that the expectation value <r2) <* (N + 3/2)]. Therefore, as the nucleus
deforms, the nuclear matter approaches this outer orbit. Since the nuclear
force is attractive, the energy of a particle in this orbit decreases. The effect is
obviously less for a particle in a lower oscillator shell that is already closer to
(or inside) the bulk of the nucleus when it is spherical.
In the opposite limit of large deformation, the 1 • s and I 2 terms in Eqs. 7.2
and 7.5 are negligible and the Hamiltonian simply reduces to an anisotropic
harmonic oscillator whose form shows that the motion clearly separates into
independent oscillations in the z direction and in the xy plane. Therefore the
number of quanta in these directions, nz and (n^ + ny), separately become good
quantum numbers. The eigenvalues of the one dimensional harmonic oscilla-
tor with quanta n. are simplyft<».(«i+1/2). This gives the familiar result for an
isotropic three-dimensional harmonic oscillator that E = nco(N + 3/2) where
N - «j + «2 + ny Thus, in the present case of large S, the eigenvalues of the
anisotropic harmonic oscillator of Eq. 7.2 go asymptotically to

Since // is independent of the angle $ around the z axis, the Hamiltonian

corresponding to Eq. 7.10 is invariant with respect to rotations about the z axis.
Therefore the z-projection of both the orbital and spin angular momenta of a
 The Deformed Shell Model or Nilsson Model 259

particle must be constants of the motion. As we have stated, these quantum

numbers—the eigenvalues of the operators /z and sz—are commonly denoted
by A and £ while their sum, the projection of the total angular momentum on
the symmetry axis, is indicated by K. The asymptotic energies E(nx, ny, nz) can
then be more conveniently expressed in terms of the quantum numbers
K[NntA] of the familiar Nilsson orbit notation.
Asymptotically, these energies are dependent on nt and independent of A.
The separation according to nz, that is, according to the extent of the motion in
the z direction or perpendicular to it, simply reflects the point made at the
beginning of this chapter that, since the nuclear force is attractive, equatorial
orbits will be favored and polar orbits unfavored in energy. The independence
of A occurs for large 8 because the terms in 1 • s and I 2 are negligible. Since
K = A ± Z, this independence of A becomes an independence of K for large 5.
This is exactly opposite to the small deformation limit.
In general, a given value of nz will have a number of degenerate states that
can be specified by A, taking on the values (N-nz), (N-n-2), (N-n^-4), ...0 or
1. For finite deformation where the 1 • s and I 2 terms cannot be ignored, the
eigenvalues will also split according to the value of A and, therefore, of K.
The asymptotic separation of the Nilsson diagram for large deformation
according to nz and the approximate independence of A or AT are surprisingly
little known, but can easily be seen in the Nilsson diagram for large deforma-
tions. As evident in Fig. 7.4 for large e, the lowest-lying orbits have nt = N, N-l,
while in midshell nz = 1-2 orbits predominate and near the end of a shell the
n = 0 and n = 1 orbits are collected. The independence of A or K is illustrated
nicely by the nearly degenerate and parallel orbits pairs 7/2~[503] and
9/2-[505], 3/2-[512] and l/2-[510], or 3/2+[422] and l/2+[420].
A nice empirical verification of the separability of the motion into compo-
nents along and perpendicular to the symmetry axis comes from the properties
of certain orbits differing by ±2 in their principle quantum number TV. In
principle, the Nilsson Hamiltonian (specifically the I 2 and I • s terms) can
couple states with AN = ±2, although these couplings are normally neglected
since such states are separated by two oscillator shells (=10 MeV). For large
deformations, however, the sensitivity of the energies to nz leads to the
phenomenon that steeply upsloping orbits from oscillator shell Nmay eventu-
ally cross steeply downsloping orbits from the N + 2 shell. These orbits will
have small and large values of n , respectively. An example of such AW = 2
orbit pairs are the 3/2+[402] and 3/2+[651] orbits. A priori, their mixing might
be expected to be large in the near crossing region. However, that mixing has
been empirically deduced from single nucleon transfer cross sections. The
extracted interaction matrix elements are typically only =50-100 keV. Such
small coupling matrix elements between states with very different distribution
of quanta in the z and xy directions points to the approximate validity of the
separation of motion in these two perpendicular directions. It is worth noting,
however, that the presence of other deformation components, such as hexade-
capole (/J4) shapes, can greatly increase AA7 = 2 mixing.
To recapitulate some of the preceding points, we see two limiting situations
of the Nilsson scheme. For small deformations S, the energies are approxi-
260 Collectivity, Phase Transitions, Deformation

mately given by Eq. 7.8. They are linear in S, j remains an approximately good
quantum number (the configuration mixing is still small), and the orbits are
separated principally by their K quantum numbers. For large deformations,
the energies (Eq. 7.10) are again linear in 5 (recall that the ft), are linear in S).
The slopes, however, now depend on n and the energies separate according to
the distribution of motion along and perpendicular to the z axis. For interme-
diate deformations, a transition between these two coupling schemes takes
place.

7.2 Examples
Having discussed the Nilsson model both physically and formally, we can now
turn to its application to odd mass deformed nuclei. Actually, this works in
much the same way as the shell model for the single-particle excitations of
spherical nuclei. The principle difference lies in the degeneracy of the orbits.
In the shell model, an orbit; can contain 2j +1 nucleons. In the Nilsson model,
the degeneracy is broken according to the orbit orientation, or K value, and
each Nilsson orbit can contain only two nucleons, corresponding to the two
ways (±K) in which the nucleon can orbit the nucleus (clockwise or counter-
clockwise). Neglecting pairing for a moment, in a deformed region the Nilsson
orbits are sequentially filled, two protons and neutrons to each, until the last
odd nucleon is placed. This defines the ground state. Excited single-particle
excitations can be obtained two ways, either by raising the last odd nucleon to
a higher orbit, thereby changing its Nilsson quantum numbers, or by lifting a
nucleon from one of the filled orbits to the last orbit, completing a pair of
nucleons in the latter and leaving a hole below the Fermi surface. One
therefore expects to have a sequence of intrinsic excitations whose energies
and quantum numbers can be simply read off from the Nilsson diagram once
the deformation is specified.
Here, in effect, is the major difference between the spherical shell model
and the Nilsson model: N = 105 corresponds to 21 holes relative to the magic
number 126. A typical shell model calculation would diagonalize some resid-
ual interaction among 21 neutrons in the 82-126 shell and the complexity
would be enormous. By switching to a deformed basis, the Nilsson model
regains a "single-particle" picture, but with deformed single-particles orbits,
each a relatively simple mixture of spherical / orbits. Multiparticle (or quasi-
particle) excitations (the equivalent of seniority v > 3 in the spherical shell
model) only begin to appear near the pairing gap at 1.5-2.0 MeV. The
deformed ansatz gives a remarkable simplification.
We recall and emphasize here that the Nilsson wave function is only a
specification of the orbital motion of the nucleons in a body fixed coordinate
system: the full specification of the wave function requires a consideration of
the rotational behavior. This is absolutely crucial for an understanding of the
structure of odd mass deformed nuclei and, indeed, for a deeper understand-
ing of the Nilsson model itself, as well as its testing and application to real
nuclei. We shall turn to the rotational motion shortly.
 The Deformed Shell Model or Nilsson Model 261

It is nevertheless useful at this point to indicate how well and simply the
Nilsson model can be applied to deformed nuclei by way of a brief example or
two. Consider the nucleus 177Hf with 72 protons and 105 neutrons. All the
protons will be paired off to total angular momentum zero and, at least for the
low-lying single-particle excitations, can be ignored. The same applies to the
first 104 neutrons. Simple counting in the Nilsson scheme for e~ 0.3 shows that
the 105th neutron will enter the 7/2~[514] orbit. We therefore expect that the
ground state of 177Hf will be 7/2". (Actually, this is not so trivial: we have
implicitly assumed that the lowest angular momentum will be equal to the K
value for a given orbit. While this is generally true, it is not always the case,
especially when strong Coriolis effects are present. This is a question that must
be dealt with when we consider the rotational motion of an odd nucleus in
more detail. For the moment we accept this assumption.) A low-lying excited
state can clearly be formed by lifting the last neutron to the 9/2+[624] orbit,
giving a 9/2+ state. Similarly, one of the two nucleons in the 5/2~[512] orbit may
be raised into the 7/2"[514] orbit leaving a hole with spin 5/2~. Other low-lying
excitations should correspond to the l/2-[521], 7/2-[514], and 7/2~[503] orbits at
appropriate energies. A partial empirical level scheme for 177Hf is given in Fig.
7.5, showing the bandhead levels corresponding to each intrinsic Nilsson
excitation. It corresponds rather well with our predictions. If we now go to
179
Hf, we would expect the ground state to be 9/2^624] with the 7/2-[514] an
excited (hole) state. Moreover, all the excitations that were above the Fermi
surface in 177Hf will now decrease in energy while those that were below the
Fermi surface will increase in excitation energy. Comparisons of the two-level
schemes in Fig. 7.5, which uses the convention that particle excitations are
shown on the right and hole excitations on left, reveals exactly this behavior.
In general, as one sequences through a series of isotopes (or isotones, if one is

Fig. 7.5. limpirical bandheads of intrinsic Nilsson excitations in 177 Hf, 17'JHf. Particle (hole) states
arc on the right (left).
262 Collectivity, Phase Transitions, Deformation

Fig. 7.6. Systematics of some Nilsson orbit excitation energies in the rare earth region. Hole states
are given negative energies (extracted from Bunker, 1971).

dealing with odd proton nuclei), the energy of a given Nilsson orbit should
descend along the right-hand side of the "V." At some point it should become
the ground state, or at least occur very low in the spectrum, then increase in
energy along the left arm of the "V." At least approximately, the behavior
exemplified by the systematics in Fig. 7.6 (here hole energies are shown as
negative values) is typically observed. Deviations from it can be due to
changes in deformation across such a sequence (we have implicitly assumed a
constant deformation), or to shifts in the relative positions of the Nilsson orbits
from effects such as higher order deformation components (hexadecapole
deformations), or to Coriolis effects.
A nearby nucleus that shows one such case is 183W, whose level scheme will
be discussed at great length in the next chapter and is illustrated in Fig. 8.1.
Simple counting would suggest that the ground state is 7/2-[503], with low-ly-
ing l/2-[510] and 3/2~[512] particle excitations. Yet, the empirical level scheme
shows that the latter two orbits are near the ground state and the 7/2" [503]
occurs at a few hundred keV excitation energy. An explanation of this will be
given in Chapter 8.
We have seen that it is as easy in the Nilsson model as in the shell model to
determine the expected order of single-particle excitations and their energies
and to deduce, virtually by inspection, an anticipated level scheme. Though
this seems a trivial exercise, one should not lose sight of the fact that by
considering a deformed shell model potential, one is able to account instantly
 The Deformed Shell Model or Nilsson Model 263

for the low-lying levels of literally hundreds of deformed odd mass nuclei,
ranging from A ~ 20 to the actinides. There have been innumerable tests of this
model over the last three decades and it has proved capable of correlating a
vast amount of data, particularly when some rather simple refinements (pri-
marily Coriolis mixing and hexadecapole deformations) are incorporated.

7.3 Prolate and Oblate Shapes

It is interesting to break the discussion at this point to discuss an extremely
basic question that is seldom alluded to but is now easy to answer. It is an
empirical fact that the vast majority of deformed nuclei are prolate rather than
oblate in their ground states. The only candidates for oblate nuclei are those
in which either N or Z is near the very end of a major shell (e .g., Hg). However,
the Nilsson diagram can be applied equally on the oblate side.
Although we will not discuss it explicitly, the derivation of the model for
oblate shapes should be self-evident by now. Here the "core" nucleus is disc
shaped, and the lowest energy orbits will be polar with high K values. The
sequence of levels will be more or less inverted relative to the prolate case. For
example (see the sketch in Fig. 7.3), the i13/2 shell model state will again split
into seven orbits in order of increasing energy K - 13/2,11/2,9/2 1/2. For
small | f) |, the energies are again linear in K1, and hence the spacings between
different K levels are proportional to K.
Although the empirical preference for prolate shapes is well known, it is
much less understood why it should be so. Years ago, this issue was discussed
with some frequency, but since has been ignored without having received a
satisfactory or at least well-known explanation. However, a simple under-
standing actually involves only two considerations, namely the specific se-
quence of single-particle spherical; shells and the relative angular orientations
of different K orbits in the Nilsson model.
A casual inspection of the Nilsson diagram seems to show no preference for
oblate or prolate shapes. Energy-favored downward sloping orbits appear on
both sides. Furthermore, if we imagine a spherical closed shell nucleus with a
single valence nucleon (see Fig. 3.1), the resulting disk-like orbital "ring"
would appear to be oblate. There is no elongation along a symmetry axis but
an orbitalp/ane superimposed on the spherical core.
Why, therefore, are most nuclei, and in particular those at the beginning of
major shells, prolate? Consider a single / shell such as im. We recall from an
earlier discussion (see Table 7.1) that the orientations of the orbital planes
change very little for low K values, but increase rapidly for higher K. Thus,
while on the oblate side, a single K = 13/2 orbit may descend as rapidly in
energy as its K = 1/2 partner on the prolate side, there are several prolate orbits
that have comparable, strongly downsloping energies. In contrast, only a
couple of orbits will be steeply downsloping on the oblate side. We saw before
that the "switch" to upsloping should occur at K = 9/2. When several valence
nucleons are present, as there must be for the nucleus to be deformed, a
summation over their energies will favor a prolate deformation.
264 Collectivity, Phase Transitions, Deformation

This is almost, but not quite, the desired explanation. One other ingredient
is necessary. To see this, consider the sequence of spherical / values (see Fig.
3.2, or Fig. 7.4 for ft = 0). Suppose that, instead of this order, the lowest orbit
had been p1/2, pM, or even f^ instead of a higher j orbit. The preceding
argument would then have had little weight. Upsloping orbits would soon be
encountered for both prolate and oblate shapes. It is because the lowest orbits
after a shell closure have relatively high j, with many K values, that a distinc-
tion between the oblate and prolate behavior can be made and a preference for
prolate deformations can develop. Thus, the second key feature is the modifi-
cation of the shell model potential to include components that favor lower
energy for higher / and; orbits. The nearly universal preference for prolate
shapes in nuclei stems from the specific radial shape of the shell model central
potential that is intermediate between harmonic oscillator and square well,
which favors large I values, and from the properties of the sine (the orbit
inclinations as a function of K). Once again, we have an example of how a very
simple but physically intuitive appreciation of the structure of a given model
leads to important results even without detailed calculation.

7.4 Interplay of NHsson Structure and Rotational Motion

We have discussed the structure of the Nilsson wave functions as linear
combinations of single; shell model wave functions with expansion coefficients
C. and have seen a number of examples of such wave functions. It cannot be
emphasized too strongly that these are wave functions in the body-fixed
system, that is, the nonrotating nucleus. In this system, / is clearly not a good
quantum number. However, the nucleus exists in space and the total angular
momentum J must be a constant or the motion. The projection of this angular
momentum K, (and the expectation value (/'} of the single-particle angular
momentum) on the nuclear symmetry axis are also good quantum numbers.
The Nilsson wave function is known as an intrinsic state or a state of excitation
of an isolated body, which in this case is the deformed nucleus. The real
nuclear states are combinations of this intrinsic motion and a superimposed
rotational motion of the core. Phrased another way, the Nilsson wave function
does not possess a fixed angular momentum J; rather, it can be projected onto
states of many different angular momenta. It is because of this seemingly
abstract idea that a particle in a given Nilsson orbit actually gives rise not to a
single state in a deformed nucleus (as would be the case for a single shell model
particle in a given orbit j), but rather to a set of states comprising what is known
as a rotational band.
A proper understanding of the interplay of this rotational motion and the
intrinsic motion leads to a much deeper understanding of the wave functions
for odd mass deformed nuclei, of the actual nature of the rotational motion
involved, of the reasons why single nucleon transfer reactions are such power-
ful probes of Nilsson model wave functions, and of the effects of the Coriolis
interaction.
The same approach used to obtain the first order rotational energy expres-
 The Deformed Shell Model or Nilsson Model 265

Fig. 7.7. Angular momentum diagram for an odd mass deformed nucleus. Note: this figure can be
very misleading—see text and Fig. 7.9.

sion in even-even nuclei can also be used for odd mass nuclei. If the odd
nucleon, considered for a moment to be in a single; orbit, does not polarize the
even-even core, then the total angular momentum results from the vector sum
of the core rotation and the odd particle angular momentum. This is illustrated
in Fig. 7.7. We now start with the same rotational Hamiltonian as before and
obtain, using the notation for the different angular momenta given in Fig. 7.7,

We can convert this to a more useful form by defining the familiar raising and
lowering operators

Simple multiplication of these operators gives J + j + J j + = 2(JJj + J2j2).

Therefore

and hence

Replacing these operators with their eigenvalues where possible and using the
fact that, for low-lying states, both J3 and/, have the same projection K, gives
266 Collectivity, Phase Transitions, Deformation

or

where

The first term on the right in Eq. 7.14 is identical to the rotational energy
expression for an even-even nucleus. It is clear from Fig. 7.7 that this must be
the case, since this figure would collapse to that of an even mass rotor if there
were no single-particle angular momentum;. The other terms in Eq. 7.14 arise
specifically from the presence of the odd particle and are intimately connected
with the coupling between rotational and particle degrees of freedom. Even
the resemblance of the first term to the symmetric top formula is fundamen-
tally misleading. We shall see momentarily that the simple picture illustrated
in Fig. 7.7 conceals some important physical effects, and that the rotational
motion is not as simple as commonly believed, but an alternate picture that is
nearly as simple will allow us to retrieve Eq. 7.14 in a transparent, elegant way
that will disclose a much different understanding of rotational motion in odd
mass nuclei.
The third and fourth terms in Eq. 7.14 involve the / structure of the de-
formed single-particle wave function (Nilsson wave function). Whatever
value (f) takes, both it and K should be constant within a rotational band
(neglecting band mixing). The last term is called the Coriolis interaction
because its effects are very similar to the classical Coriolis force acting on any
rotating macroscopic body. The Coriolis interaction has important conse-
quences in both even and odd deformed nuclei and will be extensively dis-
cussed later. We will show then that, in first order, its effects on energies simply
correspond to a change in the WLI.
To this order then, the energy levels of a given rotational band in an odd
mass nucleus should behave as ft2I2IJ(J + 1). In Fig. 7.8, we illustrate a number
of examples of rotational bands in heavy nuclei and show that this simple
formula works remarkably well. Also included in the figure are two examples
where it clearly fails to provide even a reasonable first-order estimate. One of
these involves a K - 1/2 band (1/2~[521] in 169Er) that we shall later see incor-
porates a special (diagonal) Coriolis interaction. The other is a "band" ("h11/2"
in 133La) that appears to be partly "upside down" (e.g., E(ll/2-) « E(5/2-),
E(7/2~)) and unrelated to the kind of structure that we have been examining. It
too involves especially strong Coriolis effects and will be discussed later. Here,
we wish to raise, and resolve, an apparent paradox that arises from Fig. 7.7
(and that is intimately related to this type of inverted structure).
Equation 7.11 states that the total angular momentum results from the
vector combination of the rotational angular momentum R and the particle
angular momentum/, that is J = R + j. The situation was sketched in Fig. 7.7,
which is a simplification since the Nilsson wave function contains, in general, a
linear combination of functions of different/. That is not the point. Consider,
 The Deformed Shell Model or Nilsson Model 267

Fig. 7.8. Rotational bands in some deformed odd-mass nuclei. The dots give the rotational energy
predictions from Eq. 7.14 after normalization to the first two levels. Dots for the energies of the
K = 1/2,1/2~[521] band in 169Er are omitted since they require a "decoupling parameter" term (to
be discussed in Chapter 8). As discussed later in the text, the rightmost "band" is so Coriolis mixed
that no single Nilsson label is possible.

for simplicity, the one case in which a single / value does nearly characterize the
Nilsson wave functions—the unique parity orbits. To be specific, let us take a
Nilsson wave function such as that for the l/2+[660] orbit from the i13/2 neutron
shell in the rare earth region. (For convenience we neglect Coriolis mixing.)
Now, we have seen examples of rotational bands with spins J = K, K + 1,
K + 2,... whose energies vary approximately as J(J + 1), or in this case, a
sequence with J = 1/2,3/2,5/2,7/2,.... The common view (Fig. 7.7) of this band
as consisting of a single particle in the l/2+[660] orbit coupled to a sequence of
successively faster core rotations is seriously in error. To see this, recall that we
have taken a simple case where the Nilsson wave function consists of only one
; value,; = 13/2. This is therefore the only single-particle angular momentum
in the system. This in turn implies that any total angular momentum other than
/ = 13/2 must incorporate angular momentum from another source. That
source can only be the rotational motion. Thus, as in Fig. 7.7, one can imagine
a J = 17/2 state obtained by coupling a; = 13/2 single-particle angular momen-
tum to a core rotational angular momentum R = 2. Similarly, J = 9/2,5/2 states
could be formed by the antiparallel coupling of / = 13/2 and R = 2, R - 4,
respectively. The only energies in the system are the Nilsson energy, which is
constant (independent of J), and the rotational energy, fi.2/2IR2. Hence, the
J = 9/2 state (with R = 2) should have higher energy than the/ = 13/2 state (with
R = 0). The J = 5/2 and 1/2 states should be expected still higher. Moreover, the
energy difference £(9/2) - £(13/2) should equal £ R _ 2 - £R=0, or in other words,
£2\ of the neighboring even-even core nucleus; £(5/2) - £(13/2) should equal
£4^ , and so on. This picture leads to "upside down" rotational bands with the
268 Collectivity, Phase Transitions, Deformation

lower spin states lying higher than the state with J=/. Such band structures do
indeed exist (as we noted in discussing the right-most band in Fig. 7.8). They
have recently become highly interesting as a particular manifestation of im-
portant Coriolis effects in high-spin states. However, they are not the normally
observed situation and are certainly not consistent with most of the empirical
rotational bands shown in Fig. 7.8. Clearly, there is something wrong with this
picture.
A clue to a more accurate understanding of the rotational motion begins by
recalling that we are dealing with an axially symmetric deformed nucleus. This
means that any orientation of the angular momentum vector j with respect to
the symmetry axis z that maintains a projection K is indistinguishable from any
other orientation and therefore is equally likely: the angular momentum
vector j is free to precess around the z axis. Figure 7.7 showed only one
particular orientation of this angular momentum vector—that corresponding
to the smallest possible value of \R\, since j and R are aligned. In contrast,
imagine that the angular momentum j were rotated 180° to that shown in the
figure so that it lay in the plane of the page but pointed downward, below the
z axis. This situation is depicted on the left in Fig. 7.9. The amount of core
rotation required to produce a final total angular momentum J would clearly
be much larger. If we extend this idea to other angles of the angular momen-
tum vector j, then, as j processes, R will point in a continually varying direction
and I R \ will take on a constantly changing series of values.

Fig. 7.9. (Left) Angular momentum diagram for an odd-mass deformed nucleus. This is a more
rigorous version of Fig. 7.7 that incorporates important refinements. (Right) Path (circle)
followed by the tip of the j vector with time as it precesses about the z axis (point C). This is an
"end view" of the time dependence of the diagram on the left. I am grateful to D. D. Warner for
this figure.
 The Deformed Shell Model or Nilsson Model 269

Consider again our example of a/ = 13/2 particle in a K - 1/2 orbit. For

7 = 1/2, / points nearly along the z axis, and I R \ remains roughly constant in
magnitude, although not in direction, as j precesses, at a value 1 R\~6. On the
other hand, for / = 13/2, \R\ takes on values ranging from 0 to 12 as i precesses.
Since the rotational Hamiltonian, £ 2 /2/R 2 , is quadratic in R, large ] R \ values
carry more weight than small ones. Therefore, a situation in which | R \ varies
smoothly from 0 to 12 will have, on average, higher energy than one in which
I R I is approximately constant at a value I R \ - 6, and the J = 13/2 state will
require more (rotational) energy than the / = 1/2 level.
This analysis can be made more quantitative by viewing the nucleus end on.
This situation is depicted in the side and end views in Fig. 7.9. Point C is along
the symmetry axis at a distance K from the origin. Point B is the terminus of
the fixed total angular momentum vector J, and point A marks that of the
single-particle angular momentum j. The circle in the end view shows the path
followed by the angular momentum vector j as it precesses about the symmetry
axis z. We can now calculate the expectation value of R 2 by allowing j to rotate
about the point C at a radius r. From simple trigonometry

Integrating this expression over of 0 gives

but

and

Thus

or, taking the quantum mechanical expectation values

This is exactly the same as Eq. 7.14 except for the Coriolis term that has not
been included since we have assumed that K is constant and, as we shall see,
the Coriolis interaction inherently mixes different K values. (Since we as-
sumed a single-/ wave function, (j2) has become /(/ + 1).) Note that the
assumption of constant K is equivalent to the assumption that the particle
angular momentum vector j precesses exactly about the z axis. We shall
shortly encounter a special, though not uncommon, situation in which preces-
sion is about an axis perpendicular to the z axis: clearly in such a situation it will
be K, not R, that changes continuously. This then will bring us back to the
picture in Fig. 7.7, which leads to "inverted" rotational sequences.
We therefore see that the derivation of Eq. 7.14 for the eigenvalues in a
rotational band built on a given Nilsson orbit was correct, even though the
270 Collectivity, Phase Transitions, Deformation

simple picture that is commonly used to illustrate this situation, Fig. 7.7, is too
simplistic. The more accurate view shown in Fig. 7.9 gives the same formula in
a trivial manner. It has been a constant emphasis in this book that most results
in nuclear structure physics can be derived, at least semi-quantitatively, by
very simple, often intuitive, analyses. This example warns us that such an
approach cannot be careless handwaving, but must accurately reflect the
correct underlying physics.
The most important conclusion from the present analysis is the recognition
that the rotational motion in a deformed odd mass nucleus is far from simple.
Not only does the magnitude of the rotational angular momentum I R \ vary
with time but its direction in space also changes. In the / = 1/2 state, for
example, the nucleus at times rotates clockwise about the y axis, at other times
counter clockwise about this axis, and at still other times about a continuously
varying axis in the xy plane. The rate of rotation is relatively constant
corresponding to \R\ =6, In the 7=13/2 state, on the other hand, the nucleus
is at times stationary (R = 0) while at others it rotates at frequencies varying
from values corresponding to R = 0 to those for | R \ = 12! We can see that the
rotational motion is really a complex, time dependent, variation that includes
not only a true rotational component but also a kind of tumbling motion. The
fact that the principle term in Eq. 7.14 (i.e., the term that depends on /, the
others being constant) has the same form as in the symmetric top or in the
rotational energy expression for an even-even nucleus, is really almost an
accidental result of the particular combination of the processing single-particle
motion and the varying core rotational motion needed to produce a constant
total momentum J*.
For typical (nonunique parity) Nilsson wave functions, several / values
commonly appear and the "rotational" motion is even more complex. From
the formal standpoint, one need not worry about this since, as we saw in
deriving Eq. 7.14, as long as the proper vector character of the angular
momenta are taken into account the correct results always emerge.
There is one other important point implicit to this discussion. It should be
clear by now that while many values of I R \ contribute to the wave function for
a state of spin / in a rotational band built on a given Nilsson orbit, the special
value R = 0 can only occur if the Nilsson wave function y/Mfa contains an
amplitude 0,—that is, only if it contains an amplitude for the single nucleon in
the orbit with j = J. For any other / value, the state with total angular
momentum J can be constructed only by incorporating some rotational mo-
tion. As we shall see, this has important consequences for single nucleon
transfer reactions. We recall that for a given oscillator shell N, ;mai = N +1/2
(e.g.,;mai = 11/2 for N = 5). It is thus clear that all states with J>N + l/2 must
have R * 0.
Having discussed the basic Nilsson model and the basic rotational motion in
odd mass deformed nuclei, we turn in the next chapter to a detailed discussion
of specific tests of and refinements to the model, with emphasis on the crucial
and pervasive effects of Coriolis coupling.
'The author is grateful to D. D. Warner, with whom this analysis was worked out.
 8
NILSSON MODEL:
APPLICATIONS AND R E F I N E M E N T S

8.1 Single Nucleon Transfer Reactions

The description of odd mass deformed nuclei in terms of Nilsson orbits and
their configuration-mixed nonspherical wave functions has been an extremely
successful model for over three decades. One of its most appealing features is
that it is extremely easy to test empirically and to measure the detailed shell
model (/) composition of individual Nilsson wave functions. As we shall see
momentarily, single nucleon transfer reactions provide a direct and specific
measure of each successive component in the Nilsson wave functions. This
remarkable property stems from a particularly simple feature of the interplay
of rotational and single-particle motion in producing final states of given total
angular momentum /.
Although we shall defer our discussion of experimental probes of nuclear
structure to a later chapter, an exception will be made for single nucleon
transfer reactions leading to deformed odd mass nuclei, because they are so
intimately linked with the basic structure of the Nilsson model itself. The
underlying reason stems from a point about the rotational motion brought out
at the end of the previous chapter. Since J = R + j, a component with R = 0 in
a state of spin J can only occur if the Nilsson wave function i//= EC 0. contains
an amplitude for the single-particle angular momentum / = J. If we could
somehow sample the R = 0 components of a sequence of states in a rotational
band, we would be sampling the successive C amplitudes in the Nilsson wave
function for the underlying intrinsic state. Single nucleon transfer reactions do
just that.
We will pursue this point in a moment, but first it is interesting to consider
how and under what conditions such reactions occur, as well as some of the
experimental considerations applicable to them. We shall do this for the (d, p)
reaction, but the same reasoning applies to other light-ion single nucleon
transfer reactions such as (d, t), (3He, a), (a, t), and so on. First, consider a
schematic diagram of such a reaction as was shown in Fig. 3.3. An incident
nucleon, in this example a deuteron, passes near a target nucleus. As it
experiences the nuclear force (we neglect Coulomb effects) several things may
occur. The deuteron may simply scatter from the nucleus elastically or
inelastically (producing excited states). It could be absorbed by the target
nucleus producing a nucleus A' = A + 2 in a highly excited "compound" nuclear
state. It could attract one nucleon, such as a neutron, from the target nucleus

271
272 Collectivity, Phase Transitions, Deformation

and emerge from the collision as a triton, leaving behind a residual nucleus
with A' = A-l. This is a (d, t) reaction. Or, a neutron could be stripped off and
enter an orbit around the target nucleus, producing a final nucleus with mass
A' = A +1, a (d, p) reaction. These reactions can be experimentally selected by
detecting the outgoing particle and identifying it. This can be done with a
number of different techniques we will not discuss here. Some utilize the
different magnetic rigidities of the outgoing particles. Others exploit the
dependence on the mass and charge of the ratio of the energy loss AE in a thin
detector to the total energy.
In any case, we assume we have identified an outgoing proton, thereby
"tagging" a (d, p) reaction event. This reaction is not necessarily a single step
process. It could be accompanied by inelastic scattering or Coulomb excita-
tion. Or it could be the result of a compound nuclear reaction in which the
deuteron was first fully absorbed. To select appropriate experimental condi-
tions to favor a simple single step process, we consider some of the parameters
describing the reaction process of Fig. 3.3. If the closest distance of the
incident projectile from the target nucleus is large, the interaction is weak, and
stripping occurs with low probability. When scattering occurs it will not be at
large angles. Two-step processes involve a product of such single-step ampli-
tudes, and will be negligible. For a close collision (distance of closest ap-
proach), the nuclear interaction and hence the scattering angle are much
greater. The probability of a single event occurring is much larger but so is that
for multistep processes. The optimum situation of large, direct, reaction cross
sections but small multistep amplitudes occurs for an intermediate angle
corresponding to a "grazing" collision. This discussion is highly qualitative
since quantum mechanical interference effects lead to significant oscillations
of o(d, p) with 9. Nonetheless, for targets with A > 100 and typical incident
deuteron energies Ed~ 10-15 MeV, most (d, p) experiments emphasize detec-
tion at angles 40° < 6< 125°. Similar considerations apply for heavier projec-
tiles, except that, for a given bombarding energy per nucleon, the heavier
projectile brings in more angular momentum and is more likely to excite final
states involving larger angular momentum transfer.
With this brief digression finished, we turn to consider the nuclear structure
information obtainable from single-step reactions such as (d, p). We consider
the population of members of a rotational band built on some Nilsson orbit in
a deformed odd mass final nucleus. Since the reaction is single step, sequential
processes, such as transfer followed by inelastic scattering, are eliminated by
the choice of experimental conditions. Indeed, by definition—the only thing
that can occur is that a single nucleon can be transferred to a given, quantized,
empty valence orbit (Nilsson wave function t//Nils). In particular, the process
cannot induce any rotation of the target nucleus. Since no rotational notion
can be imparted, it follows that the probability of populating a state with a
given total angular momentum J must be proportional to the probability, C2._,,
for a shell model single-particle wave function 0 ; in the Nilsson wave
function. Thus, even though the intrinsic wave function for each state in the
rotational band is identical, the (d, p) cross section for populating each
 Nilsson Model: Applications and Refinements 273

successive state successively samples a specific component of the Nilsson wave

function. The cross sections for populating the / = K, K + 1, K + 2, K + 3,...
states directly give C2K, C2K+l, C2K+2, C2K+3 in v^Nfls Since the sequence of C
values is characteristic of each specific Nilsson wave function, the pattern or
cross sections is an identifying signature of a particular orbit and is commonly
called a fingerprint pattern. Much of the study of Nilsson orbits in heavy nuclei
is grounded in this basic property of single nucleon transfer sections.
Note the interesting point brought out in the last chapter that since
-/max = N + 1/2 for a given shell, single-step single-nucleon transfer reactions
can never populate states with J > N + 1/2. All components in the wave
functions of such states must have R *• 0. If such states are populated, it is
immediate evidence either for multistep processes or for wave function admix-
tures from higher major shells (e.g., AN= 2 mixing that can be induced by large
hexadecapole deformation components).
The formal expression for the single nucleon transfer cross section to a
specific state of spin / of a rotational band in a deformed nucleus built on a
Nilsson orbit /, is given by

where we have included a pairing factor denoted by P2, which we shall discuss
in a moment. (The quantity (C P)2 is analogous to the spectroscopic factor
(see chapters 2 and 3) for spherical nuclei.) This formula is extremely well
known and has been extensively used for two decades in probing the structure
of Nilsson wave functions; strangely, its simple origins described earlier are
often only vaguely understood, and the power of (d, p), (d, t), and other single-
particle reactions in elucidating Nilsson structure often seems almost magical.
Before discussing this equation in relation to the empirically deduced
structure of various Nilsson orbits in typical deformed nuclei, we must add a
few more comments on the other factors appearing in it. In a formal derivation
of Eq. 8.1, there must occur an integral linking the initial and final states

giving the overlap of the initial deuteron-(A)-nucleon even-even target system

with the final proton-(A + 1) nucleon odd mass system. This matrix element
involves the internal structure of the incoming and outgoing projectiles and
the degree to which the final nucleus looks like the target plus a neutron in a
specific orbit. It is usually assumed for simplicity that the latter point is
satisfied: (\j/ms(A+l) \ ^n\jff^(A)}2 is unity. This is really just the single-step
process assumption. The former aspect concerning the projectile/ejectile
structure is absorbed into the arbitrary normalization constant N in Eq. 8.1.
The reaction process also depends on kinematic effects (Ed, 6, etc.). These
kinematic factors are included in the function 0((0), which can be calculated by
standard DWBA techniques. Typically, 0( (0) has a diffractive oscillatory
pattern that is a function of 9. The specific extent and locations of maxima and
minima are functions of the transferred angular momentum /. Starting from an
Fig. 8.1. Level scheme for 183W. Levels without Nilsson assignments are given at top center (Casten, 1972). Particle (hole) excitations are on the right (left).
 Nilsson Model: Applications and Refinements 275

even-even nucleus, the final angular momentum J = j = I ± 1/2 and the final
state parity is n= (-1)'. In principle, a measurement of the angular distribution
of the outgoing particles can provide information on the J* values of various
final states. In practice, this information is somewhat unreliable in deformed
nuclei, and measurements are typically made at only two or three angles: the
ratio of the cross sections at these angles provides at least a qualitative guide to
the transferred orbital angular momentum /. On account of the centrifugal
barrier, it should not be surprising that, for low-energy, light projectiles, the
population of higher / values is inhibited: the cross sections decrease with
increasing /(and therefore J). Generally, it turns out that the angular distribu-
tions for small / values are somewhat forward peaked, while those for large /
transfers are backward peaked. Therefore, a ratio such as a(125°)/o"(600)
increases with transferred orbital angular momentum /. The cross sections
<t>t(9) also have a dependence on the reaction Q value (the difference in
incoming and outgoing projectile energies). The Q value is easily deduced
from the known nucleon separation energies. For example, for a (d, p)
reaction

where B.E.(d) is the deuteron binding energy 2.23 MeV. Since S(n) = 5-8 MeV
in heavy nuclei, Q(d, p) is typically positive for low Ex and decreases as Ex
increases.
As noted, the use of reactions that carry more momentum into the system
such as (3He, a), favors high / transfers. Therefore, the ratio of populations of
a given state in (a, 3He) and (d, p), a(a, 3He)/o(d, p), can also serve as a
"meter" for the transferred angular momentum /. Indeed, at back angles, it
singles out the highest// values accessible (e.g.,/ = 13/2 in the odd neutron rare
earth nuclei: see the following discussion (Fig. 8.9)).
The factor P2 is U2 for a stripping reaction such as (d, p), and V2 for a pickup
reaction such as (d, t). It represents the probability that the single nucleon
orbit involved is initially either empty or filled, respectively. It is reasonable
that this factor is present. In a (d, p) or (d, t) reaction, a given orbit can be
populated only to the extent that it is initially empty or full, respectively. Thus
(d, p) tends to populate orbits above the Fermi surface, while (d, t) populates
orbits below the Fermi surface most intensely.
Now we can turn to the extraction of specific nuclear structure information
from these reactions. It is easiest to show this using a specific example.
Consider the final nucleus 183W whose level scheme is shown in Fig. 8.1 with the
states arranged according to Nilsson assignment and rotational band in the
same format as in Fig. 7.5 for 177>179Hf. Figure 8.2 shows (d, p) and (d, t) spectra
leading to 183W, while Table 8.1 summarizes the measured cross sections (at
90°) for those negative parity states that were assigned to specific Nilsson
states. Table 8.2 gives similar information for 185W, We assume that the
"kinematic" factors Afy, (Q) are known so that the cross sections may be used to
extract empirical values of OP for the ith band.
Look at Fig. 8.2, bearing in mind the strong / dependence of the DWBA
276 Collectivity, Phase Transitions, Deformation

Fig. 8.Z (d, p) and (d, t) spectra leading to 183W. The peaks are labeled by the Nilsson assignments
(Casten,1972).

cross sections. Typically <j(d, p) drops by an order of magnitude as / changes

from / = 1 to / = 5. Thus, large cross sections may not imply large C2Pi values
and vice versa.
Extracting the C' coefficients and the structure of each band now only
requires an estimate of the pairing factors t/ or V. The simplest way to do this
 Nilsson Model: Applications and Refinements 277

Table 8.1 Comparison of unperturbed, Coriolis coupled and experimental cross sections in 183W
(6 =90°)*
C^d.p)^/^ o(d,t) pblsr
State Unper- Per- Experi- Unper- Per- Experi-
turbed turbed mental turbed turbed mental
9/29/2~[505] 64 65 25 1.8 1.2
7/27/21503] 357 364 284 38 48 72
9/2 7.0 8.0 =3 0.6 1.3 =1
3/23/2~[512] 138 208 131 29 60 58
5/2 249 184 96 48 34 35
7/2 30 48 87 5.8 17 44
9/2 6.0 3.9 4 1.0 1.0 <1
11/2 0.2 0.4 4 0.0 0.1 1
1/2 1/2~[510] 0.3 0.4 8 0.2 0.6 5
3/2 334 269 264 195 179 150
5/2 98 163 202 53 65 103
7/2 33 18 11 18 15 14
9/2 3.3 5.5 11 1.6 1.6 7
5/25/2-[512] 0.0 0.0 6 0.2 0.2 19
7/2 17 7.9 29 269 217 237
9/2 1.2 1.2 3 14 25 18
11/2 0.1 0.0 1.3 0.8 4
l/21/2-[521] 25 15 15 280 280 364
3/2 7.0 2.8 3 128 114 54
5/2 5.1 4.3 86 87 66
7/2 4.3 1.8 71 64 69
9/2 0.7 0.6 10 10 8
7/27/2-[514] 2.0 3.8 16 40 81 91
912 2.6 3.0 3 45 35 29
11/2 0.0 0.1 <4 0.8 1.3 15
•Casten, 1972.

is to assume a Fermi energy, and the simplest assumption here is to assume that
it coincides with the energy of the Nilsson orbit that forms the ground state in
the odd mass nucleus. For a reasonable choice of the gap parameter A (typi-
cally 0.75-1 MeV), it is easy to solve the quasi-particle Eq . 5. 23, to obtain U
and Vas a function of excitation energy. However, the Fermi energy need not
coincide exactly with any specific Nilsson orbit. For example, in 183W the
ground state is the l/2^[ 510] orbit (see Fig. 7.4) and the 3/2~[512] orbit occurs
at an excitation energy of approximately 200 keV, while in 185W the order is
reversed but the two bands occur within =20 keV each other. If the deforma-
tion has not changed, one cannot account for this asymmetric situation by
placing the Fermi surface at, say, the position of the 1/2^[510] orbit in 183W and
at the 3/2~[512] orbit in 185W. Given the separation of these two orbits in the
Nilsson diagram, <= 150 keV, it is clear that in 185W, the Fermi surface must be
approximately centered between these two orbits, producing (see Fig. 5.7) low
excitation energies for both. Since the U and V factors change rapidly near the
Fermi surface, such fine details of the Fermi surface location can have signifi-
cant effects on single nucleon transfer cross sections for low lying orbits.
278 Collectivity, Phase Transitions, Deformation

Table 8.2. Comparison of unperturbed, Coriolis coupled, and experimental cross sections in ^W
(0 = 90°)*
d(d,p) fjb/sr a(d,t) i^b/sr
State Unper- Per- Experi- Unper- Per- Experi-
turbed turbed mental turbed turbed mental
9/29/2-[505] 66 62 25 3.3 3.9 10
7/27/r[503] 334 341 316 58 72 154
9/2 6.4 10 1.0 2.1 =3
11/2 0.5 0.5 -6 0.1 0.2 1
3/23/2-[512] 107 30 5 59 1.7 1
5/2 195 247 301 97 173 207
7/2 23 7.2 11 12 1.0 1
9/2 4.7 6.5 11 2.1 4.5 11
11/2 0.2 0.1 0.1 0.0 =0.4
1/2 1/21510] 0.2 0.2 4 0.3 0.9 3
3/3 180 262 357 343 423 308
5/2 53 0.3 =4 92 14 11
7/2 18 37 104 32 55 99
9/2 1.8 0.0 2.7 0.1 =1
11/2 0.1 0.2 5 0.2 0.4 4
5/25/2^512] 0.0 0.0 <3 0.2 0.2 14
7/2 17 7.7 32 268 208 206
9/2 1.2 1.1 14.1 24 =20
1/2 1/2~[521] 14 14 <62 281 280 =266
3/2 6.4 2.8 129 107 =43
5/2 4.7 3.8 86 89 =20
7/2 3.9 1.6 72 60 =25
9/2 0.5 0.6 10 10 11
7/27/2"[514] 2.1 4.1 15 40 86 80
9/2 2.7 1.2 45 34 19
*Casten, 1972.

A more empirical way to extract U and V factors is from the ratio of (d, p)
and (d, t) cross sections to the same state in an odd mass nucleus of mass A.
This procedure is slightly inconsistent, since the U2 factors refer to the empti-
ness of Nilsson orbits in the target nucleus A in the (d, p) reaction while the V2
factors relevant to (d, t) refer to the orbit occupancies in its A + 2 target
nucleus. Nevertheless, this technique is widely used and is adequate for
essentially all cases of practical importance.
We could now extract empirical C values from the measured (d, p) and (d, t)
cross sections. For technical reasons, it is easiest and most common to use
theoretical sets of C coefficients to calculate theoretical cross sections and to
compare these with the measurements.
By identifying appropriate sequences of cross sections, the states in an odd
mass nucleus may be sorted into rotational hands whose Nilsson wave func-
tions can be identified. Such assignments are made in Tables 8.1 and 8.2 for
183, iss^ an(j [he theoretical and experimental cross sections are compared. In
 Nilsson Model: Applications and Refinements 279

Fig. 8.2, the deduced J values and Nilsson quantum numbers are indicated
above the corresponding peaks.
Careful inspection shows very different patterns for the different rotational
bands, justifying the term fingerprint patterns. For example, the 1/2~[510] band
has a very small cross section to the l/2~ state, large cross sections to the 3/2-
and 5/2- levels, and smaller cross sections thereafter. The 7/2-[503] band has a
large cross section only for the 7/2" state. For the positive parity levels,
essentially only the 13/2+ states are populated. (Note that we have not
attached specific Nilsson quantum numbers to some of the latter levels. The
reason will be clear after we have discussed the strong Coriolis mixing between
these bands.) The U1 and V2 dependence in Eq. 8.1 is also evident in Fig. 8.2.
For the low-lying bands where V1 and V2 are roughly comparable, the same
states were populated in both (d, p) and (d, t), whereas at higher energies, the
levels separate according to whether they are populated in stripping (LP) or
pick up (V2). For example, the 1/2 1/2" [521] and 7/2 5/2- [512] hole states are
stronger than the ground band states in (d, t), but much weaker in (d, p).
Tables 8.1 and 8.2 are well worth careful inspection. Although there are
small differences in detail, especially for weaker states, the characteristic
fingerprint patterns are often observed experimentally. Examples are the
5/2-[512] and l/2-[510] bands in 183W and the 7/2-[503] band in 185W. Indeed,
these fingerprint patterns are often the technique used to identify the specific
Nilsson orbits in the first place. The reader should not minimize the impressive
successes of such a simple model, many of whose predictions can be antici-
pated without calculation despite the presence of perhaps dozens of valence
nucleons.
Nevertheless, while qualitative patterns emulate the data, the detailed
predictions often disagree substantially with the experimental results. Ex-
amples are the 5/2 3/2-[512], 5/21/2-[510], 3/21/2-[521], and 7/2 7/2-[514] states
in both nuclei. There are differences of nearly an order of magnitude in the
cross sections for populating certain corresponding states in the two nuclei.
For example, in (d, p) in 183W, the 5/21/2~[510] state is strongly populated while
the 5/2 3/2-[512] state is weak; in 185W, it is just the opposite.
Both of these phenomena are striking manifestations of the importance of
the Coriolis interaction in odd mass nuclei. As we shall see, we can greatly
improve the predicted cross sections if we take this residual interaction into
account. The importance of the Coriolis interaction goes far beyond the
question of sorting out difficulties with single nucleon transfer reactions. It is
especially important for high-spin states in both odd and even mass nuclei, and
has been shown to lead to a new coupling scheme—the so-called rotation
aligned coupling scheme characterized by "decoupled" bands in many nuclei,
and by the backbending phenomenon. It is in fact difficult to overestimate its
significance in understanding odd mass deformed nuclei. We turn now to a
systematic treatment of the Coriolis interaction with emphasis on its physical
origin, its principal effects, and a simplified discussion of some easy ways to
estimate its effects by inspection.
280 Collectivity, Phase Transitions, Deformation

8.2 The Coriolis Interaction in Deformed Nuclei

The origin of the Coriolis interaction has already been seen in Eq. 7.13, where
the rotational energy expression for a single-particle coupled to a deformed
rotor contains a term -h2/2I (J j + JJ ) where J± acts on the total angular
momentum and j+ on the particle angular momentum. This term is an interac-
tion between the rotational and single-particle motion and has physical effects
similar to that of the Coriolis force on a classic rotating body. The simple
properties of operators such as J±, j± are discussed in any standard quantum
mechanics text. They serve as raising and lowering operators for the z
projections, K and Q, of the total and single-particle angular momenta. Their
matrix elements are (equating K and Q and calling both K):

Note that both J+ and j decrease K while both J and j+ increase K.

The physical nature of the Coriolis interaction is easy to see. It is analogous
to the classic effect that occurs in any rotating body. Consider the analogy of
a projectile traveling northward from the equator on the earth's surface,
illustrated in Fig. 8.3. A projectile launched at the equator initially travels
eastward at a rate given by the rotational speed of the earth at the equator.
Since the earth's circumference at higher latitudes is less, an observer at a
northern latitude has a smaller rotational velocity. Tb this observer, the

Kg. 8.3. Coriolis effect in a rotating system resulting from the dependence of rotational velocities
on "latitude."
 Nilsson Model: Applications and Refinements 281

projectile that was fired directly northward will appear to be deflected toward
the east. This is the Coriolis effect, and while it is sometimes called a fictitious
or apparent force, it has real physical effects. (It accounts for the fact that river
banks tend to be eroded more on the right (facing downstream) side in the
northern hemisphere.) We see from this illustration that the Coriolis interac-
tion effectively tilts the orbit relative to the equator. If we now picture the
orientation of the angular momentum vector perpendicular to the orbit in-
stead of the orbit itself, the Coriolis effect is equivalent to a change in its
projection onto the equator. Thus it is understandable that, in the nuclear case,
the Coriolis interaction alters the projection of the angular momentum K on
the symmetry axis, admixing different K values. Another way of looking at
this that will be useful later is to recall that K is only a good quantum number
if the nuclear potential is axially symmetric. Therefore, the Coriolis interac-
tion effectively introduces small amounts of axial asymmetry as it mixes K
values.
We now evaluate the Coriolis matrix element explicitly. We consider two
intrinsic Nilsson states characterized by K and K +1. Since J is a good quantum
number, we can replace J± with its eigenvalue. This cannot be done with j
because of the configuration (j) mixing in Nilsson wave functions, although it
can be done approximately for the unique parity orbits where one / value
dominates. Noting that only one of the two terms in J + j + J j+ gives a
nonvanishing result, we get

where the symbols K, K +1 in the j matrix element are a shorthand for the two
Nilsson wave functions and the effects of pairing are included in the factor
(C/jUJj + V^V^). The pairing factor has the general effect of reducing the
Coriolis matrix elements since its maximum value is unity. The reduction is
least for orbits in similar positions relative to the Fermi surface (then (7, = U2
and Vl = V2, hence (£/,£/2 + VjV2) = U,2 + V,2 = 1). It is least for orbits laying on
far opposite sides (then £/; - V2 and U2» Vv so (U^U2 + V^V2) - (U^ + U2V2~)
and one factor in each term is small). For diagonal Coriolis matrix elements,
the factor is obviously unity.
For a single j shell, the j matrix element is

This is approximately correct for unique parity orbits. For an arbitrary Nilsson
wave function, terms like this occur for each/, so that, in general,

where the CK and C *+1 coefficients are those describing the specific Nilsson
wave functions. An interesting limiting case occurs for /, / » K, if / is
approximately a good quantum number. This applies for low # orbits from the
high-spin unique parity states or for high spin, low K states generally if ft is
small (little configuration mixing). Then Eqs. 8.4 and 8.6 give
282 Collectivity, Phase Transitions, Deformation

This is a general upper limit on the strength of the Coriolis interaction. For
typical inertial parameters (say 15 keV for rare earth nuclei) and say,
J,;' =11/2, this attains = 400 keV! Typical spacings between Nilsson orbits are
« 150 keV. Coriolis mixing is not necessarily a minor perturbation!
Since the Coriolis matrix elements change K by AK = ±1, it is generally a
nondiagonal interaction. However, it has an important diagonal matrix ele-
ment that contributes to certain rotational energies. For K = 1/2, the symmet-
rization of the wave function gives rise to terms with K = ±1/2, allowing a
diagonal AK= 1 contribution to the energies from the cross terms. Substituting
Eqs 8.4 and 8.6 in Eq. 7.14, we get for the rotational energies for K -1/2 bands
including Coriolis mixing

where a is the well-known decoupling parameter given by substituting

K = -l/2inEq.8.6

The two phase factors in Eqs. 8.8 and 8.9 come from the symmetrization of the
wave functions (e.g., Eq. 6.10). The phase factor in Eq. 8.8, (-1)J+1/2, means that
the contribution to the rotational energies from the Coriolis interaction alter-
nates in sign with /.
It is clear from Eq. 8.9 that a can be either positive or negative. The
behavior of E(J) in Eq. 8.8 as a function of a is shown in Fig. 8.4. If a < 0, states
with spins 3/2,7/2,11/2... are lowered in energy while the alternate spin states
are raised. For a > 0, the opposite situation occurs. On account of the factor
(/ +1/2) the effect grows with spin. On account of the factor (/ +1/2) the effect,
on average, increases for heavy nuclei (jm <* W). For I a I =1, Eq. 8.8 shows
that the levels occur in degenerate pairs. If a = -1, the J - 3/2 state coincides
with the J = 1/2 level. Similarly, the (5/2,7/2) and (9/2,11/2) pairs are degener-
ate. Ifa = +l, the degenerate pairs are (3/2,5/2),(7/2,9/2),.... For |a| >l,the
level order within a rotational band is no longer monotonic in spin. Clearly, the
typical rotational spacings can be so severely perturbed as to obscure the
normal J(J +1) spacings and even the ordering of different spin states. As we
shall see, these effects propagate via nondiagonal Coriolis mixing, and affect
many bands with K * 1/2. We can now understand one of the anomalous
rotational spacings and sequences in Fig. 7.8, specifically that for the l/2~ [521]
band. This is a K = 1/2 band with its decoupling parameter close to unity.
For an arbitrary Nilsson wave function, many terms can appear in Eq. 8.9
for a. Frequently these terms (each carrying a phase) largely cancel and the
resultant a values are rather small, typically less than unity. However, very
large a values can be obtained if the wave function is dominated by few terms
with high; values. The classic example of this is the unique parity orbits for
 Nilsson Model: Applications and Refinements 283

Fig. 8.4. Dependence of rotational energies (7) on the decouph'ng parameter a. The dashed lines
are for | a I = 1 (based on Preston, 1975).

which(j}~N+1/2 (e.g.,; = 13/2 in theN = 6shell), and the wave functions are
nearly pure in/. For these special orbits, Eq. 8.9 gives a = (-1)N(N + 1). For
example, a ~ 1 for the l/2+[660] orbit. This enormous value so perturbs the
normal spacing that the 13/2+ level is among the lowest-lying levels in the
rotational band. For h2/2I = 15 keV, the decoupling term is ~ -750 keV! In
general, the sign of a is always such that the J = N +1/2 level (e.g., J = 13/2 for
the ijM shell) is lowered.
The reader may recall an apparent paradox in the order and spacing of
rotational energies in odd mass nuclei that was discussed in Chapter 7. The
simplest view led to the notion that rotational bands should be "upside down."
We showed that the "normal" order was regained when the precession of/
around the symmetry axis was considered. We also pointed out that in some
cases an "upside down" pattern does in fact occur. We have just encountered
that case where large Coriolis effects in unique parity orbits upset the mono-
tonic order of rotational energies with /. Having gone to great lengths to
explain away this paradox in Chapter 7, why does it now appear in the data? In
284 Collectivity, Phase Transitions, Deformation

other words, what happened to the precession argument? As we shall see later,
the physical difference here is that K is no longer a good quantum number
(±1/2 values are admixed), so the precession need not be about the symmetry
axis: the Coriolis force, by mixing K values, forces the angular momentum
vector to switch back and forth (precess) about the rotation axis instead. Thus,
j and R are nearly parallel or antiparallel, so now 7=13/2 does correspond to
| / ? | = 0 , / = 9to R | = 2, and so on.
One often reads that a is called the decoupling parameter because it repre-
sents a decoupling of the rotational and single-particle motion. It is now easy
to see why this name is appropriate. For a = +1, the energy differences of the
1/2, 5/2, 9/2,... states are exactly the same as those between the 0+, 2+, 4+,...
states of the even-even core: the odd particle acts like a spectator to the
rotation. Moreover, it is clear from the comment just made that changes in
core rotation have little effect on the orientation of J = (R +j) when a is large
(i.e., when | ;' | is large, Klj is small, so sin0= Klj is small, and J is nearly aligned
along the rotation axis). Extrapolating to very large a values in Fig. 8.4 shows
that the J = 13/2 state will lie lowest, followed by the 9/2 level (13/2 - 2), and
then by the 5/2 level (13/2-4). The alternate spin levels are much higher. Thus
the rotational energies of alternate / values (with \J-j\ even) are nearly para-
bolic in | J-j\ where /is the dominant; of the unique parity orbit. Sequential
states differ mainly in R: the rotational motion is effectively decoupled from
that of the odd particle. We shall encounter a related but even more dramatic
effect later as a consequence of the nondiagonal Coriolis interaction.
Turning now to these nondiagonal Coriolis effects, there are two significant
observable effects. One concerns rotational energies and stems from a propa-
gation of the highly perturbed level order in the K = 1/2 band to higher K bands
via successive AK = 1 Coriolis mixings. Precisely because of the large decou-
pling parameters, this is most important for unique parity states and, as we
shall see, accounts for their importance in high-spin studies where the (7 + 1/2)
factor in Eq. 8.8 becomes crucial. The other effect occurs in "normal"
rotational bands (especially in their impact on single nucleon transfer cross
sections). We shall discuss this first.
Consider the admixture of two bands as shown in Fig. 8.5. Recalling our
discussion in Chapter 1 of two-state mixing and the fact that the Coriolis
mixing increases with /, the perturbed energies will behave as illustrated. It is
easy to show that, to first order, the effect of the Coriolis interaction is to
decrease the effective inertial parameter, h2/2I, for the lower band and to
increase it for the higher band.
Equation 1.6 gives the energy shifts of the two interacting states relative to
the unperturbed spacing for a given spin /

where the last step assumes small mixing (Eq. 1.12). Thus AE^ °c V1^.
Isolating the spin dependence Eq. 8.4 gives
 Nilsson Model: Applications and Refinements 285

Fig. 8.5. Illustration of the changes in effective rotational parameters, h2/2I, resulting from two-
state Coriolis mixing.

which is just the rotational energy expression (J, K dependent parts). The
Coriolis interaction merely alters the effective rotational spacings. The lower
band is compressed, the upper one expanded. This simple result breaks down
for very large or multistate mixing, but even in these cases gives a useful
framework.
In general, there can be many low-lying Nilsson bands with assorted K
values; thus a realistic Coriolis mixing calculation will be a multistate diag-
onalization. The admixed wave functions can be written

where the at are the mixing amplitudes that depend on/.

286 Collectivity, Phase Transitions, Deformation

In this way, for example, a predominantly K = 3/2 band may contain

admixtures of K = 7/2 through the intermediary of K = 5/2. We shall see later
that this is especially important for the unique parity case. Although such
mixing can be rather complicated, one can almost always estimate the effects
of a full multistate diagonalization rather accurately by carrying out a se-
quence of two-state mixing calculations. We shall see examples of this shortly.
At this stage it is important to discuss the actual magnitudes of the Coriolis
mixing matrix elements. The quantities h2!2I, (J-K)(J+K+l), and the U, V
factors are easy to estimate, as we have done. They are dependent on the
orbits involved but not on their detailed structure (Nilsson wave functions).
However, <^Mb(^01 i I W(^-+1)) depends explicitly on the structure of the
states included, as evident in Eq. 8.6.
As with the decoupling parameter, the nondiagonal j matrix elements for
the unique parity orbits are both large and particularly simple to calculate,
since one ;' term in Eq. 8.6 dominates. For example, for j8 = 0.23 and typical
Nilsson parameters K and /z, the Coriolis matrix element of j connecting
the l/2+[660] and 3/2*[651] Nilsson orbits is = 6.6. If we had assumed the wave
functions consisted only of the; = 13/2 component, Eq. 8.6 gives (K\}_\ K+l} ~
N + 1, which is the maximum possible value for any matrix element of; in a
given shell. Thus, to a good approximation, the unique parity Coriolis matrix
elements can be estimated using the single; approximation, and moreover,
they have very nearly the maximum possible values.
Table 8.3 provides a number of other examples of off-diagonal j matrix
elements in the N = 5 shell. The reader may easily derive these numbers by
applying Eq. 8.6, using the wave functions given earlier in Table 7.2. Inspec-
tion of Table 8.3 reveals two global features. First, Coriolis matrix elements for
nonunique parity orbits are considerably smaller. Second, their relative values
can differ by more than an order of magnitude.
In considering a given level scheme and attempting to determine whether
Coriolis mixing effects will be important, it is often useful to be able to
estimate, without calculation, the approximate magnitude for Coriolis matrix
elements. A very simple rule allows one to do this. For nonunique parity
orbits, Coriolis matrix elements divide roughly into two classes, allowed and
nonallowed. The allowed matrix elements are those in which both n and A

Table 8.3. Theoretiral valuas {K I j I K + 1>* ]

9/2"[505] 7/2-[503] 3/2-[512] 1/2~[510] 5/2~[512] 1/2'[521] 7/2-[514]

9/2~[505] -0.973 2.847

7/2-[503] -0.973 2.858
3/2-[512] 0.951 0.045 2.546
1/2"[510] 0.951 -2.541
5/2~[512] 2.858 0.045 -1.151
1/2~[521] 2.546 -2.541
7/2~[514] 2.847 -1.151

•The Nilsson model parameters are S = 0. 2, K = 0.0637, /i = 0.42.

 Nilsson Model: Applications and Refinements 287

change by one unit, but their sum remains constant (An = -AA). (The unique
parity case is, of course, in this class.) Examples from the table are the matrix
elements between the 3/2-[512] and l/2-[521] orbits or between the 7/2-[514]
and 9/2~[505] orbits. Such allowed j matrix elements are typically on the order
of NI2 - M3. All of the others are nonallowed matrix elements and are
typically < 1. Examples of these are the matrix elements between 7/2~[514] and
5/2-[512] or l/2-[510] and 3/2- [512].
It is important to note that Coriolis mixing between orbits in the latter class
is not always negligible, especially if they have the same nz values. As we have
seen, the Nilsson diagram separates approximately according to nz values for
large deformations. Therefore, it can often occur that two orbits with identical
nz values differing in K by &.K = ±1 lay very close to each other. Their Coriolis
mixing can be large even with a small matrix element. A classic example of this
occurs for the l/2-[510] and 3/2~[512] orbits in ™-™W. We shall discuss this
case in some detail momentarily.
Having dealt with typical values of the j± matrix elements, it is useful to
develop a feeling for the absolute magnitudes of the full Coriolis matrix
elements in Eq. 8.4. For well-deformed nuclei, h2l2I» £2^/6 where E2\ is
given by a neighboring even-even nucleus. For rare earth nuclei, fi2!2I~ 14-18
keV, while for the actinides, fi 2/2I» 7 keV. The matrix elements of J± given by
the square root factor are typically 2-3 for moderate spin states, although they
can become very large for high spins. Finally, as noted, the pairing factor
becomes very small for high-lying orbits on opposite sides of the Fermi surface,
while for orbits near the Fermi surface, or for those on the same side, this
factor is typically between 0.7 and 1.0. Thus, typical nonunique parity allowed
Coriolis matrix elements in the rare earth region are roughly
VCor= (16)(3)(1)(0.8) <= 40 keV. This estimate is only accurate to a factor of
2-3. Nonallowed Coriolis matrix elements will, of course, be less.
For unique parity orbits, the j matrix elements are - N. In addition, the
observed states are typically of rather high /, since single nucleon transfer
reactions preferentially populate the / = N + 1/2 states for which C. ~ 1 and
heavy ion reactions tend to feed the high-spin unique parity levels. Taking
V(./-.KV./+Ar+l) = N, the Coriolis matrix elements linking unique parity
orbits can be extremely large, typically reaching V Cor (unique parity) ~
16(6)(6)(0.8) <= 400 keV. Such matrix elements mixing states often only a
couple of hundred keV apart have enormous structural effects.
From extensive experience with Coriolis mixing calculations, it has been
found that the actual empirical matrix elements are generally about 20-50%
lower than these theoretical estimates. This conclusion emerges from com-
parisons of extensive data on level energies and single nucleon transfer cross
sections in many deformed nuclei. We will not detail this evidence here, but
one example of it is trivially evident in Fig. 8.1, which shows that the 7/2~ states
of the 7/2-[514] and 5/2~[512] bands in 183W are separated by only 70 keV. If we
consider this an isolated two-state system (an approximation good enough for
the present purposes although not for detailed calculations) and recall that
such states can never be closer than twice their mixing matrix element, then the
288 Collectivity, Phase Transitions, Deformation

Coriolis matrix element between these two states must be < 35 ke V. Since both
bands are hole excitations, the pairing factor is near unity. Using h2/2I = 18 keV
(E2f (184W) = 111 keV) and the j matrix element from Table 8.3 gives a
predicted Coriolis matrix element of 55 keV. The maximum matrix element
allowed empirically is only 65% of this. Other similar examples abound.
Despite this attenuation, Coriolis mixing effects, especially among unique
parity orbits, represent a substantial perturbation to the rotational picture and
can seldom be ignored. One final point to emphasize before considering some
actual calculations is that in weakly deformed and transitional nuclei, Coriolis
matrix elements are far larger than in well-deformed nuclei because of the
smaller moments of inertia. Matrix elements between unique parity orbits
may reach an MeV or more, and under certain circumstances may even lead to
a new coupling scheme, the so-called rotation aligned scheme we shall discuss
later.
As an example of multistate Coriolis mixing, let us consider the level
scheme of 183W shown in Fig. 8.1. In principle, a full calculation cannot neglect
the unseen bands that occur at higher energies, but in practice, one assumes
that their effects are small (at least for nonunique parity orbits). We will see
one way to estimate whether such an assumption is grossly violated. Under
this assumption, the Coriolis mixing among the negative parity bands involves
diagonalizing matrices of varying size, 2 x 2 for 7 = 1/2, up to 6 x 6 for / = 11/2.
For the positive (unique) parity states, the strength of the Coriolis mixing
precludes safely ignoring unseen bands and therefore one usually carries out a
full 7 x 7 diagonalization.
We consider first the negative parity states. We note the empirical result
(from ratios of (d, p) to (d, t) cross sections) that the Fermi surface is slightly
below the 1/2~[510] orbit. For simplicity we ignore the 9/2[505] band. (In any
case, it can only affect/ 9/2 states.) From our earlier discussion, we anticipate
that the principle mixing effects will occur between the l/2~[521]-3/2~[512] and
5/2-[512]-7/2-[503] pairs. That this is not quite true highlights the other factors
that must be taken into account in practical situations. Although the matrix
element connecting the l/2-[510] and 3/2~[512] orbits is rather small (=1), they
lay so close to each other that the mixing is substantial. Likewise, the
"forbidden" matrix element between the 5/2-[512] and the 7/2~[514] orbits
(=1.1), strongly admixes these close-lying bands. In contrast, despite the large
j matrix element between the 5/2~[512] hole orbit and the particle excitation
7/2~[503], the pairing factor substantially reduces the overall matrix element.
The one exception to the simple rule given above for estimating Coriolis
matrix elements among these bands occurs for the 1/2~[510] and 1/2~[521] pair:
the supposedly forbidden j matrix element has a value = 2.5. Although these
bands are nearly an MeV apart, the coupling between them is nonnegligible.
Thus, the principle Coriolis admixtures will be between the 1/2~[510] and
3/2-[512] bands, the 5/2~[512] and 7/2-[514] bands, and the l/2-[510] and
l/2-[521] bands. Second-order mixtures of, say, the l/2-[510] into the 5/2-[512]
band, will be very small. Thus, a rather good simulation of the full diagonali-
zation should be obtainable by considering sequential two-state mixing of the
 Nilsson Model: Applications and Refinements 289

Table 8.4. Calculated mixing amplitudes (a. x 100) for the 1/2' states inm 185W

K4Nn\] ,4 Ex 7/21503] 3/2-[512] 1[/21510] 5/2-[512] :t/2-[521] 7/21514]

7/2-[503] 185 244 +100 — — +5 — —
183 453 +100 — — +5 —
—
3/2-[512] 185 174 — +91 +42 — +2 —
183 412 — -96 +26 — -8 —
l/2-[510] 185 334 -41 +90 -11 —
— —
183 207 — +27 +96 -5 —
—
5/2-[512] 185 986 -5 — — +98 — -19
183 1002 -5 +98 — -17
_
— — — —
l/2-[521] 185 1335 +6 -9 -99
183 1265 — -6 -7 — -100
—
7/2-[514] 185 1058 +1 — — -19 -98
—
183 1072 +1 — — -17 — -99
*Casten, 1972.

preceding three pairs. As an example, consider the 7/2* states with a Coriolis
attenuation factor 0.7. With H2I2I = 18 keV and a pairing factor of 0.9, we
obtain =41 keV for the full l/2-[510]-3/2-[512] Coriolis matrix element. The
final spacing of the 7/2 l/2-[510] and 7/2 3/2-[512] states is 205 keV. Working
backwards in Fig. 1.7, we see that R must be rather large and therefore the
energy shift induced in each state by the mixing is a small fraction (=0.05) of
their unperturbed spacing. Reducing the full separation by 10% to estimate
the unperturbed splitting gives R ~ 3.8. Another application of Fig. 1.7 or Eq.
1.8 gives the admixed wave functions as y ("7/2 l/2-[510]") =
(0.97) 7/21/2-[510] + (0.24) 7/2 3/2-[512] and the orthogonal combination.
For the 7/2 1/2- [510J-7/2 l/2-[521] mixing we take a pairing factor of 0.6,
giving the full Coriolis matrix element of = 88 keV. AE^ = 1058 keV, so we can
use the final spacings to obtain R »12. The admixed wave functions are
y ("7/21/2-[510]") = (0.99) 7/21/2-[510] + (0.08) 7/21/2-[521] and the orthogo-
nal combination. Finally, for 7/2 3/2-[512] and 7/2 l/2-[521], the pairing factor
is ~ 0.5. Calculations again give R ~ 12, and final wave functions of
V/-("7/2 3/2-[512]") = (0.99) 7/2 3/2~[512] + (0.08) 7/21/2[521], and the orthogo-
nal combination. In all three cases the signs of the amplitudes are arbitrary.
We can test these estimates by reference to the detailed wave functions
resulting from a full diagonalization given in Table 8.4. The three admixtures
just calculated are 1/2- [510]-3/2-[512] = 0.24, l/2-[510]-l/2-[521] = 0.08, and
3/2-[512]-l/2-[521] = 0.08. The exact calculations give 0.26,0.05 and 0.08!
We can also estimate the energy shifts. Using the same R values we get the
following results (in keV):
7/2 l/2-[510]: -12.0(3/2~[512]) - 7.2(l/2-[521]) = -19.2 keV
7/2 3/2-[512]: +12.0(l/2-[510]) - 5.8(l/2-[521]) = + 6.2 keV
7/21/2-[521]: + 7.2(l/2-[510]) + 5.8(3/2-[512]) = +13.0 keV
290 Collectivity, Phase Transitions, Deformation

where the orbits in parenthesis give the mixing partner that induced each shift.
Again, these estimates are close to the results of an exact calculation.
With these shifts, and similar ones for other / values, the 1/2"[510] band is
compressed and the 1/2~[521] and 3/2~[512] bands are expanded, reflecting the
derivation in Eq. 8.10 that, to first order, Coriolis induced energy shifts can be
absorbed into changes in ft2/27. Indeed, one clue to the presence of Coriolis
effects in empirical level schemes is unequal empirical h2/2I values (i.e., after
mixing), with larger magnitudes for the higher-lying (expanded) bands and
smaller values for the lower (compressed) states. A measure of the adequacy
of a calculation is whether the input (unperturbed) ft 2/27 values are substan-
tially closer: they should be if the deformation is the same for all excited states
and small microscopic "blocking" effects are neglected.

8.3 Coriolis Mixing and Single Nucleon Transfer Cross Sections

Strong Coriolis effects are at the heart of most current research in high-spin
states, and thus occupy a crucial role in modern nuclear structure physics.
They were first studied extensively in single nucleon transfer reactions,
however, and, although such work is not so common nowadays, it is an
appropriate starting point for our discussion since the effects of Coriolis
mixing are so dramatic and also easy to understand physically in this context.
While the energy shifts previously discussed may seem rather small and per-
haps easily negligible, such an impression is misleading because even small ad-
mixtures can have large effects on single nucleon transfer cross sections. The
expression for the cross section to a state of spin / in a given band in the
presence of Coriolis mixing is an obvious generalization of Eq. 8.1 given by

where the a.'s are the Coriolis mixing amplitudes and the sum is over the
admixed bands. Note that the sum is coherent, thus magnifying the effects.
Simple manipulations also show that the total cross section is conserved for
each spin /: that which is lost by some states must be gained by others.
Before considering the example of 183W in detail, it is useful to emphasize
how small mixing amplitudes can have significant effects. For simplicity,
assume a two-state mixing of bands with identical C coefficients and pairing
factors for some spin J. Then, if the mixing amplitude of each band in the other
is 0.22 (meaning that the amplitude for the "parent" state is still 0.975), this
gives a 50% increase in the cross section of one state and a 50% decrease in the
other [o^ (1.22)2, a2<x (0.78)2]. The two cross sections that would have been
equal without mixing now differ by a factor of three!
Another feature is evident from Eq. 8.12. If two admixed states have very
different unperturbed Ci values for some J, the state with the larger C. value
will be relatively unaffected while that with the smaller may be drastically
altered. Indeed, much of the resultant cross section may easily come from the
small admixture rather than from the parent orbit itself. To be specific,
suppose the two bands have equal pairing factors, that O =0.2 and C2 = 0.8,
 Nilsson Model: Applications and Refinements 291

and that the mutual mixing amplitudes are ±0.22. Then, assuming that the
phases are such that the cross section for the state of spin J in band 1 is
increased, the ratio of perturbed to unperturbed cross sections is =3.5 for band
1 [(0.2 + 0.22(0.8)]2 and =0.9 for band 2 [(0.8 - 0.22(0.2)]2. This is another
example (bandmixing in even nuclei was the first) of how relatively small
mixing interactions and amplitudes can lead to drastic effects on certain
observables, especially when one of the unperturbed transition rates is small
or forbidden. If the phases were reversed (which would happen if the unper-
turbed positions of the two bands were exchanged), the same analysis shows
that despite the small mixing, the cross section for band 1 would essentially
vanish while that for band 2 would increase only by about 6%. Finally, if one
C. coefficient is nearly zero, the cross section will come only from the mixing.
It will therefore be independent of the signs of the mixing amplitudes and will
always be increased by the mixing.
Simple application of Eq. 8.12 to the mixing amplitudes such as those given
in Table 8.4 for the 7/2 - states of the negative parity bands in 183W gives the
cross sections labeled "perturbed" in Table 8.1. The point of this section is
highlighted by the enormous differences between perturbed and unperturbed

Fig. 8.6. Systematics of experimental, unmixed, and Coriolis coupled (d, p) and (d, 1) cross
sections in W isotopes.
292 Collectivity, Phase Transitions, Deformation

cross sections even in cases where the mixing amplitudes of Table 8.4 are small.
For example, because of its small value of Cm, nearly all the cross section to the
1/2 l/2-[510] state stems from weak mixing with the 1/2 l/2-[521] state. The
cross sections for the 3/2~ levels of the 1/2~[510] and 3/2~[512] bands are
significantly shifted by the mixing. The same applies to the / = 5/2 and 7/2
states of these bands and to the 7/2 states of the 5/2-[512] and 7/2-[514] bands.
Table 8.2 shows similar Coriolis mixing results in 185W.
Figure 8.6 shows the systematics of some empirical and Coriolis calculated
cross sections across the odd mass W isotopes. It includes a comparison with
the unmixed cross sections. The latter are constant except for small, smooth
changes in the pairing factor P2. in Eq. 8.1. The figure highlights the changes
in single nucleon transfer cross sections brought about by the Coriolis interac-
tion as well as the dramatic shifts that can occur from one isotope to another.
This is particularly evident for the l/2-[510] and 3/2~[512] bands in 183W and
185
W. An understanding of this is obvious from our discussion of two-state
mixing in Chapter 1. As we have stated before, these two bands have
interchanged positions (energies) in these two nuclei. In 183W, the 1/2~[510]
band is the lower, ground state orbital, while the 3/2~[512] band is the first
excited intrinsic excitation. In 185W, the 3/2~[512] orbital forms the ground state
and the l/2-[510] band is close, lying just above it. Therefore, the signs of the
mixing amplitudes are inverted between 185W and 183W. In the familiar termi-
nology of first-order perturbation theory, the sign of the energy denominator
has changed from one nucleus to the other. Therefore, for those states where
the two unperturbed Nilsson wave functions have comparable C coefficients,
cross sections that were increased in 183W will be decreased in ™5W and vice
versa. For cases where the C. coefficient in one is negligible, the cross section
to that state will increase relative to the unmixed case in both nuclei. Another
example of inversion concerns the 7/2~[514] and 5/2~[512] bands, in which the
7/2~levels interchange positions between 181W and 183W. Figure 8.6 shows the
dramatic effect on the weaker cross section.
Before turning to the positive parity levels, it is worth re-emphasizing the
extremely large effects involved here. Empirical fingerprint patterns auto-
matically incorporate the effects of Coriolis mixing and can differ from those
predicted by the Nilsson model by sufficiently large quantities as to completely
obscure the identification of the bands if Coriolis mixing is not taken into
account. Moreover, the mistakes that one would make would not even
necessarily be the same in neighboring nuclei, and the systematics of the
Nilsson orbits deduced could be completely wrong.

Unique Parity States

We now consider the unique parity orbits. Here, the Coriolis mixing becomes
at once much stronger but also somewhat simpler to interpret. The reason is
that, for practical purposes, only the J = N + \I2(J = 13/2 for the odd neutron
rare earth nuclei) state is important because of the utter dominance of the
C.=N+m coefficients in the Nilsson wave functions for these orbits. Moreover,
the Coriolis matrix elements are all very similar and only directly link adjacent
 Nilsson Model: Applications and Refinements 293

orbits stemming from the same/ shell (although second-order (A/C = 2) mixing
is significant).
Before discussing practical calculations, let us take a schematic model.
Assume the Fermi surface lies below the whole group of unique parity orbits in
some nucleus. The order of their excitation energies is K = 1/2,3/2 (Af+1/2).

Fig. 8.7. Illustration of how the irregular rotational spacings in strongly decoupled K =1/2 bands
can propagate to K & 1/2 bands via Coriolis mixing.
294 Collectivity, Phase Transitions, Deformation

Fig. 8.8. (3He, a) cross sections for 183W (based on Kleinheinz, 1973).

Each mixes with the K + 1 and K - 1 member of the series. We thus have a
situation analogous to one discussed in Chapter 1, in which equally spaced
states each mix with adjacent levels. One general result is that lowest band will
be pushed much lower. Moreover, given the increase of the Coriolis matrix
elements with J, it will be severely compressed, and its wave functions will be
a complex mixture of several components with all wave function components
in phase.
The principle difference between this schematic situation and the real one
arises because of the large decoupling parameter for the K = 1/2 band. For
N = 6, this has the effect of greatly lowering the 13/2+, 9/2+, 5/2*, l/2+ states and
raising the ll/2 + , 7/2+, 3/2 + ,... states of that band prior to mixing. Consider now
the effect of Coriolis mixing on the nearby K = 3/2 band. The situation is
illustrated in Fig. 8.7. The reordering of energies in the K = 1/2 band because
of the large diagonal Coriolis effect (decoupling parameter) causes the unper-
turbed spacings between the 3/2, 7/2, and 11/2 states of the two bands to be
much larger than between the 5/2,9/2, and 13/2 states. Therefore, in the lower-
lying K = 3/2 band, the 5/2, 9/2, ..., group is shifted down substantially more
than the 3/2,7/2,..., group. The perturbed energies of the K = 3/2 band take on
an alternating pattern as well, relative to a pure .7(7+1) rotational spacing, and
appear as iflhe. K - 3/2 band had a decoupling parameter of the same sign and
slightly smaller magnitude than the K = 1/2 band. When the K = 3/2 band in
 Nilsson Model: Applications and Refinements 295

Fig. 8.9. (3He, a) spectra in W isotopes (Kleinheinz, 1973).

turn mixes with the K = 5/2 band, this "signature" is passed on in a somewhat
reduced form. In effect, the Coriolis mixing "propagates" the decoupling
parameter throughout the entire sequence of unique parity orbitals.
If the Fermi surface is below the K = 1/2 orbit, the effect is reversed. The
K = 1/2 band lies below the K = 3/2 band and therefore the closest lying pairs
are the 3/2, 7/2, and 11/2 states. Also the propagation is severely damped by
the pairing factor as one goes from hole states to particle slates. In the W
isotopes that we have been considering, the Fermi surface is near the 9/2" [624]
and ll/2+[615] orbits, and the effect of the K = 1/2 band is negligible. (This is
linked with a point we will make shortly, that strong diagonal Coriolis matrix
elements are most effective in inducing a rotation aligned coupling scheme
when the Fermi surface lies near the low K orbits.) In W, the primary
observable mixing among the unique parity orbits should be in the K = 5/2,
112,9/2, and 11/2 orbits. As contrasted with the normal parity states, here the
matrix elements and spacings are comparable and a two-state mixing calcula-
tion is hopelessly crude. The results for 183W of an explicit calculation of the
single nucleon transfer strengths (C 13/2 coefficients) to 13/2' states is shown in
the top two panels of Fig. 8.8. Since the j matrix elements arise almost solely
from they = 13/2 term in Eq. 8.6, and since the C. 13/2 coefficients all have the
same sign, the phases of the resulting wave functions are such that in any two-
296 Collectivity, Ph ase Transitions, Deformation

state mixing of these orbits, the cross section to the lower state is increased
while that to the higher is decreased. This persists in the multistate extension,
and the net effect is to transfer cross section from the higher-lying bands to the
lower ones. This is the point alluded to in the schematic model at the beginning
of this discussion. We now compare this calculation with the empirical situ-
ation.
In 183W, the lowest-lying unique parity orbit is ll/2+[615]. The 9/2+[624] and
7/2+[633] orbits are hole excitations. An ideal reaction to study these unique
parity orbits is the (3He, a) reaction, which preferentially excites higher-spin
hole levels. In the rare earth region it can almost be used as a "/ = 13/2+ meter."
Typical (3He, a) spectra, for 181-185W, are shown in Fig. 8.9. Comparison with
the (d, t) reaction in Fig. 8.2 vividly illustrates the selectivity. In the absence of
Coriolis mixing, the (3He, a) reaction to 13/2+ states in 183W should look like
that shown schematically at the top of Fig. 8.8, in which there are five nearly
equally strong peaks. (That for the K = ll/2+ is weaker than the others due to
the smaller V2 factor and the peak from the 13/2+[606] orbit is absent since
V ~ 0.) In contrast, the data show only four peaks but with the same total cross
section expected for the six unperturbed states. This illustrates both the
shifting and the descent of strength just discussed. (The empirical (3He, a)
spectrum (Fig. 8.9) for 181W shows this effect even more; only two peaks
consume nearly all the 13/2+ strength.) The bottom-most panel of Fig. 8.8
summarizes the empirical C =13/2 coefficients in bar graph form for 183W.
The second panel of Fig. 8.8 includes Coriolis mixing, and is somewhat
better than the unmixed calculations shown in the top panel. Further improve-
ment requires, as we shall now see, the introduction of hexadecapole deforma-
tions. The study of such shape components offers us an ideal situation to apply
the same kind of intuitive approach we used for the Nilsson model itself.

Hexadecapole Deformations and Unique Parity States

Although quadrupole distortions dominate heavy nuclei, hexadecapole ef-
fects are not at all negligible and, as we shall see, frequently have major impact.
It is by now well known that most heavy deformed nuclei have either positive
or negative hexadecapole deformations superimposed on their quadrupole
distortions. Figure 6.10 illustrated the effect of hexadecapole deformations on
the nuclear shape for both signs of £4. For £4 > 0, the nucleus takes on a so-
called pin cushion or barrel shape, while for £4 < 0, it is "clover leafed."
Remarkably, we can now understand, without calculation, the origin of the
shape components, their expected systematics, and their effects on Nilsson
energies, on Coriolis mixing, and on single nucleon transfer cross sections. To
illustrate the usefulness of this approach, we shall carry out the following
discussion without any formal derivations. We will then compare our under-
standing with the results of actual calculations.
It is obvious that creating a positive (e4 > 0) hexadecapole shape requires
the occupation of orbits situated at large radii relative to the bulk of the
nuclear matter in orientations roughly 45° to the equatorial plane, so that the
"corners" of the mass distributions will be filled in. In a given shell, the orbits
 Nilsson Model: Applications and Refinements 297

with the largest radii are the unique parity orbits since they stem from the next
higher oscillator shell. From the relation sin0 ~ Klj, we see that 6 = 45°
corresponds to Klj ~ 0.7. For the im neutron orbit in the rare earth region, this
gives K ~ 9/2. Thus, the largest contributions to an increasing positive
hexadecapole deformation in this region occurs when the 9/2+ [624] orbit is
filling near the Fermi surface. Simple counting in the Nilsson scheme shows
that this occurs around A ~ 180.
At the other extreme, analogous reasoning shows that negative e4 values are
favored when very low or very high K i13/2 orbits are filling. The former occur
near the beginning of the deformed region at neutron number N ~ 92. The
latter occur near the end of the shell and serve to reduce the positive e4 values
for A > 180. The expected e4 systematics should therefore consist of large
negative values at the beginning of the deformed region that decrease in
absolute value with mass, cross zero, turn positive, and maximize around the
W isotopes, followed by a rather rapid decrease towards zero as the 208Pb
closed shell is approached. This is exactly the systematics observed empiri-
cally, as shown in Fig. 8.10.
[A technical point is worth mentioning here. In this discussion, we have
used e4 as the deformation parameter. Figure 8.10 is expressed in terms of j84.
In Nilsson's original paper, the principle discussion was carried out in terms of
deformations /32 and $,. The disadvantage of these parameters when discussing
AN = 2 mixing and hexadecapole deformations is that, even for fa = 0, there
will be finite AN = 2 mixing. In contrast, the e representation discussed in an
appendix to Nilsson's original article, was designed so that A/V - 2 mixing
vanishes when £4 = 0. The relation between e2, £4 and j32, j94 is complex and
coupled. Either set of deformation parameters may be converted into the
other by the use of Fig. 9 of the article by S. G. Nilsson et al (1969): However,

Fig. 8.10. Systematics of/3 4 in the rare earth nuclei.

298 Collectivity, Phase Transitions, Deformation

one should note that there is a mistake in this figure and that its proper use
requires the reversal in the sign of e4. Very crudely, e4 —/?4.]
We can go one step further. Since the orbit inclination 6 changes slowly for
low K, there will be more low-angled orbits (9 < 45°) than orbits near 45°.
Therefore, negative e4 deformations should predominate and the "crossing
point" to positive values should occur past midshell. This is also seen in Fig.
8.10.
The principle effect of hexadecapole deformations on the Nilsson wave
functions is to admix components with AN = ±2. Thus, the N = 6 im Nilsson
orbits will now contain some components from the N = 4, 8 shells, and the
N = 5 normal parity orbits will contain contributions from TV = 3,7. Normally,
this AN = 2 mixing is miniscule because of the large energy separation of
oscillator shells. However, inspection of the characteristic form of the Nilsson
diagram (downsloping early, upsloping late), shows that there are a few
isolated regions where steeply downsloping unique parity orbits from one shell
(e.g., i^) intersect upsloping orbits from the next lower shell (here, N = 4). If
the nuclei in such regions have large e4, then substantial AN = 2 mixing can
occur. An example of such interacting orbits was discussed earlier in Chapter
7 in the context of a comment on the relative purity of the nz quantum number
at large deformations.
Thus far we have discussed the origin of hexadecapole deformations, their
systematics, and their relatively minor effects on Nilsson wave functions. It
remains to discuss their enormous impact on Coriolis mixing and single
nucleon transfer cross sections. This impact arises mostly from the effect of e4
on Nilsson energies of orbits that can Coriolis mix. It is easy to see what the
main effects will be. Consider, for example, a large positive e4 and the i13/2 or-
bits. It is obvious that both equatorial (K = 1/2, 3/2,5/2) and polar (K = 13/2
and perhaps 11/2) orbits will be, on average, further from the nuclear matter
than for e4 = 0, and therefore their energies will increase. The mid-A" orbits
(K = 7/2, 9/2) will be closer to the nuclear matter and their energies will
decrease. Hence the overall effect will be a compression of the energy
separations from K = 1/2 to 7/2 or 9/2. Moreover, this compression will become
more extreme as e4 increases. In fact one can imagine sufficiently large e4
values that some of these K orbits may actually cross and interchange their
relative positions. Figure 8.11 shows an explicit calculation of the i13/2 energies
for fixed e2 as a function of e4. All these features appear. There is a compres-
sion, and even a crossing, near e4 ~ 0.1. The envelop of the energies can easily
be compressed by a factor of two and, therefore, the already large Coriolis
mixing among the unique parity orbits will increase still further (the Coriolis
matrix elements themselves will not substantially change).
Recall from our discussion of cross sections to 13+/2 states in the W isotopes
that Coriolis mixing calculations produce some improvement in the predic-
tions, but that significant discrepancies remain. The mixing casts some cross
section from higher lying levels into the lower ones. With the increased
Coriolis mixing that now occurs with a large positive e,, this effect will be
exaggerated, as shown in the third panel in Fig. 8.8, where we see much better
agreement with the empirical cross sections.
 Nilsson Model: Applications and Refinements 299

Fig. 8.11. Effect of hexadecapole deformations on i 13/2 Nilsson energies.

One of the beauties of the Nilsson model is its easy extendibility. We see
here an excellent example where the basic model predictions are in strong
disagreement with the data but where simple and physically reasonable refine-
ments easily remove most of the discrepancies, and thereby show both the
usefulness of the model as a starting point and also the absolutely crucial need
to incorporate certain of these extensions. The particular case we have been
considering is historically interesting as well: the large changes in im energies,
Coriolis mixing, and (3He, a) cross sections as e4 varies from 0 to 0.06 provided
the first definitive evidence for large hexadecapole deformations in the odd W
isotopes. Another interesting point is that while e4 values of zero and 0.06 can
be easily distinguished in this way, the approximate constancy of the envelope
of K = 1/2 to 7/2 orbits from e4 = 0.06-0.16 precludes a further refinement in the
actual £4 values.
There is one other consequence of large hexadecapole deformations that
should be mentioned. We have been discussing permanent or static hexadeca-
pole shape components. However, it is also possible that the nuclear potential
energy surface may be "soft" in £4, and that this will lead to hexadecapole
vibrations, just as softness in p and 7 leads to ft and 7 vibrations. It is not
surprising that the heavy even-even rare earth nuclei, especially the Os
isotopes, display rather low-lying (~1 MeV) K = 4 bands that have been
interpreted in terms of hexadecapole vibrations by Baker and co-workers. Of
300 Collectivity, Phase Transitions, Deformation

course, such K = 4 intrinsic excitations can also be thought of as double 7

vibrations: the particular states in the Os isotopes appear to be mixtures of
both modes, and interestingly, their 7 decay seems to pick up the two-phonon
character while their single nucleon transfer cross sections reveal the hexade-
capole aspect. This is a nice illustration of how a complementarity of experi-
mental approaches can highlight different features of nuclear excitations.

8.4 Coriolis Effects at Higher Spins

Our rather brief treatment of the unique parity orbits does not even begin to
hint at their importance in modern nuclear structure physics, especially at high
spin. Indeed, the unique parity states are the most thoroughly studied of high-
spin levels and are nearly always used as the first testing ground for new
theoretical ideas. Some of the reasons for this should be obvious from the
preceding discussion. Primary is the purity of the unique parity states in ;,
resulting from the shell model spin orbit interaction that separates them from
other levels of the same parity by nearly the distance between major shells.
Their properties can be calculated with extremely high reliability and more
simply than most other states. Second, since Coriolis effects are crucial to most
of the physics at high spins, the fact that these states form an isolated but
strongly admixed set enhances the ease with which one may spot the influence
of extra degrees of freedom. Colloquially speaking, these states "close under
the Coriolis interaction."
Another key feature of the unique parity states is that Coriolis effects are
largest amongst the low K orbitals, both because the matrix elements are
themselves slightly larger and also because the energy separations are small-
est. Noting that these are precisely the orbits whose angular momenta are
aligned most nearly parallel to the nuclear rotation axis (their orbital motion is
most nearly equatorial), we have three factors contributing to the develop-
ment of the so-called rotation aligned scheme to which we have alluded and to
which we now turn.

Rotation aligned coupling

This coupling scheme, illustrated in Fig. 8.12, was originally introduced to
account for apparently anomalous rotational spacings in certain orbits in odd
mass nuclei (see extreme right in Fig. 7.8). Figure 8.12 shows angular momen-
tum diagrams for the cases of large and small K. We assume for simplicity that
the Nilsson wave functions can be approximated by a single ;' value. We can
write the total Hamiltonian as H = HNils + Hrot. The Nilsson energy of a single
nucleon can be represented by HNns = e + AE(jK) where the e. are the spherical
single particle energies and the AZJ are the shifts due to the deformed
potential. Writing Eq. 7.8 for AEQK) as a constant plus a /C-dependent term,
we have Ems ~ e. + const + C8K?, and hence (neglecting the constant term)
 Nilsson Model: Applications and Refinements 301

Fig. 8.12. Diagram for strong coupling and rotation aligned or decoupled level schemes (based on
Stephens, 1975).

This is the Nilsson energy of the odd particle plus the total rotational energy.
At the top of Fig. 8.12 K is large and the K2 term dominates the VCor term.
Since the coefficient of K2 is linear in 8 and since the inertial parameter, h 2/2I,
decreases with increasing deformation, this is a situation that is valid for large
deformation and/or large K, The dominance of the K2 term implies that K is a
good quantum number (the K mixing terms are relatively small) and one has
the so-called deformation aligned or strong coupling scheme we have been dis-
cussing. Coriolis mixing effects are a small (but important) perturbation that
causes the angle of the vector; to "wobble" slightly as it precesses about the z
axis.
However, there are situations in which this coupling scheme does not occur.
An obvious one is for high spins for which \Cor <* 7; when this term dominates
the K2 term, the solutions must be approximate eigenfunctions of VCop that
correspond to a new coupling scheme in which K specifically is not a good
quantum number. One then has the situation illustrated in the lower part of
Fig. 8.12, where the particle angular momentum vector, j, precesses about an
axis perpendicular to the symmetry axis (i.e., about the rotation axis). Clearly,
A" will vary significantly and include negative values. It is now the alignment
along the rotation axis, commonly called a, that is the good projection quan-
tum number. This coupling scheme will be realized when

that is, especially for low K values. This is physically plausible since j already
points nearly along the rotation axis. Clearly, if K is large, an enormous
Coriolis interaction (extremely high J values) would be required to enforce
precession about the rotation axis, whereas for low K, such precession can
302 Collectivity, Phase Transitions, Deformation

occur at relatively low spins. Inspection of Eq. 8.14 shows that the rotation
aligned scheme can also be realized for low J if the coefficient of K2 vanishes.
Since ft2/27 ~ 1/5, it is clearly possible to choose a 5 value that satisfies this
cancellation requirement. For the A ~ 130 region, numerical estimates give
<5= 0.17 (/?=0.18). This is a rather moderate deformation and accounts for the
fact that the rotation aligned scheme often manifests itself in transitional,
moderately deformed prolate nuclei early in a shell (where the low K orbits are
filling). For larger deformations, the inertial parameter ft2/27 drops rapidly
while the Nilsson energies further split and the Coriolis effects decrease below
a critical value. In most well-deformed nuclei we see normal (strongly coupled)
rotational behavior.
The energies in the rotation aligned scheme are very easy to visualize. The
total angular momentum points essentially along the rotation axis and is
composed of the particle angular momentum j plus a core rotation R. Thus,
from Eq. 8.13

For high spin unique parity states, low K values and moderate deformations,
JJ » K and / is nearly a good quantum number. Moreover, in the rotation
aligned scheme, I / M a i . Therefore, using Eq. 8.7 and neglecting terms
independent of J, we have

This equation is simply that for a rotor of spin (/-a). But |(/-a)| = \R ,
the core rotational angular momentum! So the energies do not behave like
those of a rotor with spin /, but rather like those of the rotational core.
Moreover, the lowest energies occur for the highest alignments, a. The reader
may recall that when we derived Eq. 7.17, we set the problem up as the solution
to why rotational bands were not upside down, and why for example, the core
angular momenta R were not 0,2,4... for states with J = 13/2 and (9/2,17/2),
(5/2,21/2),..., respectively. The solution involved recognizing the variation of
I R | values that occurs when the particle angular momentum vector precesses
around the symmetry axis. We alluded to the possibility that rotational bands
with core rotational spacings did indeed exist in certain circumstances. We
now see those circumstances—when the precession is no longer about the
symmetry axis but rather about the rotation axis—so that R is nearly a constant
of the motion.
To understand the implications of Eq. 8.15, let us take as an example a
situation of maximum alignment for the rare earth nuclei where the unique
parity orbit has; = 13/2. We take a = 13/2. The energy difference £17/2 - E13/2 is
given by the energy difference between R = 2 and R = 0, that is, by the energy
of the first 2+ state in the even-even core nucleus or 6(ft2/27). This is com-
pletely different from the strongly coupled case in which £17/2- El3/2 =
ft2/2I [17/2(19/2) - 13/2(15/2)] = 64 (ft2/27). This difference is enormous, as is
the energy saving if the rotation aligned scheme is applicable. It was precisely
 Nilsson Model: Applications and Refinements 303

Fig. 8.13. Comparison of rotational spacings in Ba and La nuclei. This is the classic example of
decoupled band structure (based on Stephens, 1975).

the observation of such spin and energy sequences in moderately deformed

odd mass nuclei, such as the La isotopes, with energy spacings nearly identical
to those of the adjacent even-even Ba nuclei, that inspired the development of
the rotation aligned scheme. The remarkable La-Ba comparison is shown in
Fig. 8.13. The similarity of spacings in the odd and even mass nuclei is striking,
extending even to the way they track with neutron number.
The sequence of R = 0,2,4,6,... values leads in turn to low-lying rotational
states that differ in spin by two units (the alternate spin states are higher in
energy). This is another difference from the strongly coupled case of "normal"
rotational sequences J = K, K+l,K + 2,K + 3>,....
Note that it is also possible for R and j to be antialigned, producing total spin
values / < ;'. Thus, the state with J = 9/2 also corresponds to a core angular
momentum R = 2 and will have an identical energy to that for the J = 17/2 level.
Since the energies follow an R(R + 1) rule, and since states with spins J =j±R
have identical energies, it follows that a plot of E(f) vs. / will be a parabola
whose minimum is at or near J = j. It is possible, of course, that the alignment
along the rotation axis does not attain its maximum value. For example, if
a = 11/2, one can still apply Eq. 8.15. In this case the lowest aligned state, with
R = 0, will have J = 11/2. States with / = 7/2 and / = 15/2 will occur higher with
R = 2. This sequence of states also forms a parabola, but one lying slightly
higher than for the case of maximum alignment. Continuing this process, se-
quences of parabolas will occur: each succeeding one corresponds to the next
lower a value. The states along the lowest parabola are called the favored
states and may be either favored aligned or favored antialigned. Those on the
higer parabolas are called unfavored states and may also be aligned or an-
tialigned.
In a sense, this picture resembles that of a weak coupling model. There is,
however, a qualitative difference. If the coupling of a particle with angular
momentum j to a core state R is weak, all states with ( j - R)<J <(j + R) will
form a nearly degenerate multiplet since j and R can have any relative
orientations. Here, however, only the energies of the favored states approxi-
304 Collectivity, Phase Transitions, Deformation

mate those of a weak coupling model. The other, unfavored states are pushed
considerably higher.
Of course the exact solution for the rotation aligned scheme can be ob-
tained by explicit diagonalization of the Nilsson particle-plus-rotor Hamil-
tonian. An example taken from Stephens is shown in Fig. 8.14, for the case
where the unique parity orbit is the h11/2 and the Fermi surface is below the
K = 1/2 orbit. (Note that in all of the preceding discussion of the im orbits, the
only relevant property of the unique parity orbits was the purity in j, and
therefore nearly identical effects result for any other unique parity orbits
simply by substituting a different/. For example, if the unique parity orbit is
hn/2 rather than im, the lowest spin state will be J = 11/2 and the favored
aligned states will have spins J = 11/2,15/2, 19/2, .... This makes for a very
generally applicable scheme with close correlations from mass region to mass
region.) In Fig. 8.14, the favored aligned states are given by the thick lines, the
others as thin lines. The characteristic feature of the decoupled band emerges
clearly on the prolate side, whereas on the oblate side, the Fermi surface is
near the high K orbits, so the lowest states form a normal strongly coupled
band. For the rotation aligned scheme, the favored aligned energies remain
remarkably close to those of the core energies (which can be seen at /? = 0) even
out to relatively large deformations.
This brief summary of rotational alignment shows that it can be a rather
widely applicable phenomenon, occurring especially in moderately deformed
nuclei, where Coriolis effects are strong and deformation effects still rather
weak, whenever the Fermi surface is near the low K unique parity Nilsson
orbits. The rotation aligned scheme relies on the notion that it is energetically

Fig. 8.14. Behavior of particle-rotor level energies with [3 for unique parity levels including
Coriolis mixing. The Fermi surface is below the entire h ra set of orbits. Note the strongly
coupled pattern (&J = l,J(J+ 1) spacings) on the left and the decoupled pattern (A/= 2, £' mjn for
/ = 11/2, compressed (core) rotational spacings) on the right (Stephens, 1975).
 Nilsson Model: Applications and Refinements 305

Fig. 8.15. Empirical proton unique parity (n = +, g9/2) levels of odd mass Rb isotopes. The isotopes
span a strongly coupled (deformed) toward decoupled (weakly deformed) transition. The inset
indicates the proton orbits and the movement of the proton Fermi surface (with decreasing N),
Solid levels are favored states (dashed are unfavored) in the rotational aligned picture (based on
Tabor, 1989).

easier to achieve a given spin by combining a particle angular momentum

aligned nearly along the rotation axis with a small amount of core rotation than
it is to couple a particle angular momentum aligned elsewhere with a large core
rotation.
If we inspect a sequence of nuclei, it is possible to observe smooth transi-
tions from rotation aligned behavior when the Fermi surface is near the low K
unique parity orbits early in a shell (moderate deformation) to strongly
coupled as the Fermi surface rises. The odd Z Rb isotopes, shown in Fig. 8.15,
provide a nice example. The low N isotopes are reasonably well deformed: a
transition toward smaller deformations takes place as A' increases. The proton
Fermi surface is just below the entire g9/2 unique parity shell. Near 85Rb, a fine
decoupled structure exists: the low-lying levels form a A/ = 2 sequence, starting
at J = 9/2, that is, a fully aligned (a = 9/2) configuration. The alternate spins are
shifted quite high. (Incidentally, in modern parlance this splitting of favored
and unfavored states is known as "signature splitting".) As TV decreases, the
deformation increases, simultaneously reducing the Coriolis strength among
the g9/2 proton orbits and lowering the K = 1/2, 3/2 orbit energies so that the
Fermi energy moves into this group (see inset): a transition to a strongly
306 Collectivity, Phase Transitions, Deformation

Fig. 8.16. Plot of K (see text) against A for several energy differences in the Er nuclei (based on
Stephens, 1975).

coupled scheme ensues. By 77Rb, a nearly monotonic A/ = 1 sequence appears

although a fully developed rotational pattern has yet to emerge.
Another beautiful example of such a transition from a decoupled toward a
strongly coupled scheme occurs in the neutron deficient, near-stable Er iso-
topes. Stephens displays this transition in a way that highlights a couple of
important points. Consider a state of spin J in an odd mass nucleus. If the state
is fully aligned (a - j ) , then the core rotational angular momentum
I R | = \J-jl and the energy difference in the odd mass nucleus E(J)-E(j) =
(h2/2I) \J - j 2 = Eee( \J - j ] ) of the neighboring even-even nuclei. If the
strongly coupled limit applies, these rotational energy differences in the odd
mass nucleus will be much larger than those in the even-even nucleus for states
with J > j: £(./) - E(j) = O2/2/) [7(7 + 1) -/ 0 + 1)1- Figure 8.16 shows the data
for the lowest band based on the i13/2 orbital in the odd mass Er isotopes. The
ordinate is the ratio Rea of the energy difference E°(J + 2) - E"(j) of two states
in each odd mass nucleus divided by the energy difference E'(J + 2 - j) -
E'(J - f) taken from the data for the neighboring even-even nucleus. If the
 Nilsson Model: Applications and Refinements 307

odd mass nucleus has full rotation alignment (a= 13/2), then Reo»1 independ-
ent of spin.
If the rotational band in the odd mass nucleus is strongly coupled, the larger
spacings will lead to Rto » 1. Empirically, near 157Er Rfo » 1, but a clear
transition toward a strongly coupled limit is observed with increasing neutron
number. This is caused both by an increase in deformation, which reduces the
strength of the Coriolis interaction (VCor <* ft2/27), and by an increase in the
Fermi surface from near the low K unique parity orbits toward the mid-K
orbits. In addition, we note that the transition proceeds much more slowly for
higher spins. The energy spacing E2m- EK/2 remains very close to the rotation
aligned limit. This is simply because, as we have noted several times, the
Coriolis interaction increases with J for high spins states and therefore the
rotation aligned coupling scheme persists longer.
One last point, of some interest in terms of testing this picture of rotation
alignment and favored and unfavored states, concerns the relative role of high-
and low-spin levels. The entire picture described so far assumes a simple
axially symmetric rotational core and its coupling to the odd particle motion.
It entirely neglects any effects due to rotation-vibration coupling, axially
asymmetry, mixing with quasi-particle states, and so on. If one considers the
rather general situation of a fixed number of valence nucleons spanning a
specific set of single-particle states it is clear that, while there are many ways of
constructing low and intermediate spin states from different combinations of
the single-particle angular momenta, there is only one way of constructing the
highest spin level—by aligning all these individual particle angular momenta
along the same direction in space. Therefore, any model for this highest spin
state has the same structure, independent of the assumptions of the model.
For other high-spin (but not the highest-spin) states there will, in general, be
only a few ways of constructing them and different models may present
somewhat different, but mostly likely not very different, predictions. For low-
spin states, however, different models with different interactions may have
entirely different effects on specific subsets of states. Their energies and
structure may differ markedly from one model to another. Although the
beauty of and evidence for rotation aligned behavior is most dramatic in the
high-spin states, perhaps the most sensitive tests of such models occur in the
low-spin, unfavored, antialigned levels. Study of such levels may provide
evidence for other degrees of freedom of some importance.
We should briefly apply our understanding of Coriolis effects to even-even
nuclei. This will provide a simple understanding of the widespread phenom-
ena of backbending. Consider a specific two-quasi-particle state with both
particles in a low K unique parity orbit paired to J* = 0' and with the Fermi
surface below all the unique parity orbits. Neglect for a moment the interac-
tion between the two particles. Since the particles are in the lowest unique
parity orbits, their energies are greatly lowered due to the strong Coriolis
interaction with particles in higher K orbits. This is analogous to the situation
discussed in Chapter 1 of a set of equally spaced states with equal matrix
elements connecting adjacent levels. The lowest level is pushed down and
becomes a collective (strongly admixed) combination of amplitudes. From the
308 Collectivity, Phase Transitions, Deformation

size of the unique parity Coriolis mixing matrix elements, we have seen that
this energy lowering can be substantial. The tendency will be for each particle
to align its angular momentum with the rotation axis. Since the Coriolis
interaction grows with spin, it may well be that at some J value it becomes
energetically favorable to form a state, not from a core angular momentum

Fig. 8.17. Dlustration of the idea of crossing bands and frequencies in a plot of E(J) against J
(lower part) leading to the backbending phenomenon (top). The insets give an example, appropri-
ate to the Ce region, of how one can understand the systematics of proton (n) and neutron (v)
crossing frequencies (sketched in the small boxes) against Z and N (based on Wyss, 1989).
 Nilsson Model: Applications and Refinements 309

R = J superimposed on a spin 0+ pairing condensate, but rather by breaking the

pair of particles in a high / orbit so that their spins couple to (2; - 1) (2; is
forbidden by the Pauli principle) and coupling this angular momentum nearly
parallel to a much lower core rotational angular momentum of R ~ J - (2j' - 1).
The lower energy required because of the lower expenditure of core rotational
energy more than compensates for the energy lost in breaking the pair of
particles.
In such a case, a plot of the yrast energies against J will increases paraboli-
cally until this transition or critical angular momentum (7crit), where the sudden
drop in required R values leads to a smaller energy jump to the next higher
angular momentum. This is illustrated in Fig. 8.17 (lower), where 7crjl - 14.
The figure also shows the continuation of the "decoupled pair" band below the
critical angular momentum. That is, the configuration in which the angular
momenta of the two particles are aligned nearly along the rotation axis also
exists below spin 7: it is just not yrast. The "crossing" spin is simply that point
where this rotation aligned band begins to be favored energetically. This spin
depends on the deformation and the location of the Fermi surface relative to
the low K unique parity orbits. It therefore varies with N or Z.
Despite the radically altered structure at 7ciil, the deviation from J(J + 1)
behavior is often difficult to detect in this kind of plot. A more convenient
presentation has become standard, which is to plot the inverse of the inertial
parameter, 2Ilh2, defined for each particular spin as we will describe shortly,
against a quantity proportional to the square of the energy difference between
adjacent rotational energy states. In an ideal rotational band, E(J) - E(J - 2)
= (h2!2I) [J(J + 1) - (7 - 2)(7 - 1)] = H2I2I (47 - 2). Thus the quantity
(47 - 2)/(£(7) - £(7 + 2)) should be constant. If the moment of inertia
increases slightly with 7, due, for example, to centrifugal stretching effects,
then 211 h2 may increase gradually with 7 but, in any case, 211 h2 is empirically
defined as (47 - 2)/(£(7) - E(J - 2)). One then plots 27/ft2 against the abscissa
(E(J) - E(J - 2))2/4, which is 1/4 the square of the energy difference between
these two levels and is called (ha))2. (When plots like this are done for an odd
mass nucleus, 7 is redefined as 7 -j, where;' is the spin of the unique parity orbit
and is taken to be a good quantum number.)
Note that both ordinate and abscissa in such a plot are trivially obtained
(without ever knowing the actual excitation energies) simply from the observed
y-ray transition energies between adjacent rotational levels. At low spins, each
successive transition energy increases and 2IIH2 is more or less constant (ideal
rotational behavior) or slightly increasing. This gives a horizontal or slightly
rising line. At the crossing point, as is evident from Fig. 8.17, the level energy
increases by a smaller amount than would have been expected; that is, the
transition energy (ficai) increases less than expected, or may decrease, while
21/fi2 oc l/(£(7) - E(J - 2)) increases. A plot of 2IIH2 against (ha>)2 will exhibit
an upbend or perhaps a backbend at the crossing frequency. After this spin, a
normal rotational pattern, but with much lower R values and hence smaller
spacings, is re-established with a different "scale factor" 2Ilh2. The upper part
of Fig. 8.17 shows such a "backbending" plot for the yrast states (first crossing
310 Collectivity, Phase Transitions, Deformation

only), and aptly reflects the origin of the name backbending. The smaller
energy spacing between the J =12 and J = 14 states in the lower part becomes
the sharp backbend in the upper part. The concept of crossing frequency now
replaces that of 7crit.
One can view the reason for the upbend or backbend in two equivalent
ways, both based on the rotational energy expression (fi2/27) R(R + 1). In one,
the R values are effectively lower, as we have described. This is the view in
terms of a new aligned coupling scheme in which R suddenly decreases to take
advantage of the angular momentum gained from the aligned particle. The
other view is phrased directly in terms of Coriolis coupling in that the coupling
causes a substantial lowering of the states of the two-quasi-particle aligned
band. This lowering is larger for higher spins so that, at some spin, they cross
the "normal" levels. (Of course, there can be other origins for backbending
behavior, such as shape changes of nucleus as a whole: we do not discuss these
here.)
In a realistic situation, there will be an interaction between the ground band
and the decoupled band near the critical or crossing frequency. Just below Jctit,
the decoupled band is higher. Its interaction with a ground band will lower the
energies of the latter, effectively increasing 2I/h2, while simultaneously reduc-
ing the transition energies. This causes a slight upbend for J < 7crii: mixing
smooths out the sharply angled ideal pattern. The net effect is to lead to an S
shaped curve where the sharpness of the backbending or upbending depends
on the strength of the interaction between the two bands.
The data for many nuclei have now been accumulated and backbending is a
widespread phenomenon. As expected, it is typically observed when the
Fermi surface is near the low K orbits of the unique parity orbit. It tends to
disappear with increasing mass for a given sequence of isotopes as the Fermi
surface rises toward the higher K orbits.
How this works in practice can be illustrated by the example in Fig. 8.17.
The lower part, discussed earlier, shows a normal plot of £(/) vs. J and depicts
two crossings occurring at different E, J and crossing frequencies (h(o). The
insets show typical Nilsson diagrams for proton and neutron unique parity
orbits and indicate where the Fermi energies (EF) are in this example. We can
now understand the expected systematics of backbending in this region. As Z
increases, E*F (proton Fermi energy) increases. Moreover, the deformation ft
increases as deformation driving orbits are filled. For both reasons, the energy
required to occupy the low K unique parity orbits decreases. Hence, the
proton crossing (labeled n in the inset) occurs at lower energy and angular
momentum and hence lower hco. As N increases, the deformation decreases as
polar orbits are encountered. Hence the proton crossing frequency increases
with jV as the energies of the proton unique parity orbits rise. The neutron
crossing frequencies have an opposite behavior. They decrease with increasing
N (decreasing j3) since the energy separation to the low K orbits decreases, but
increase with Z because the deformation increases. These ideas are sketched
in the figure as crossing frequencies for protons and neutrons. While the
arguments are qualitative, they are borne out both experimentally and thco-
 Nilsson Model: Applications and Refinements 311

retically in a cranked shell model calculation in a region such as the neutron

deficient Xe-Ba-Ce nuclei, to which this example applies.
There is no reason why there should not be backbending in odd nuclei, and
this has indeed been observed. Again, the study of such states has become a
major "industry" in itself and cannot be described in any detail here. There is
one particular point deserving mention, however: the use of backbending
studies in odd mass nuclei to identify the nature of the specific particles
producing the backbend in the neighboring even-even nuclei. So far, we have
made little mention of whether or not the unique parity orbits involved are the
proton or neutron orbits. In the rare earth region, backbending can be caused,
for example, either by aligning two protons in an hn/2 orbit or two neutrons in
an im orbit. It is difficult to distinguish these two from the backbending plot
in the even nucleus itself. However, the odd nucleus can be used to resolve this
ambiguity with the technique know as odd particle blocking. The basic idea is
extremely simple. Consider an even-even nucleus in which the backbending is
caused by alignment of two-neutron quasi-particles in a specific low K unique
parity orbit. Now consider the rotational states in the neighboring odd proton
nucleus where the odd proton is in a low K unique parity orbit. As the spin

Fig. 8.18. Example of the odd particle blocking technique for intrinsic excitations in 164 165
Yb
(Riedinger, 1974).
312 Collectivity, Phase Transitions, Deformation

increases, there is no reason why two low K neutron quasi-particles cannot

align. Therefore backbending should ensue, and at approximately the same
frequency as in the neighboring even-even nucleus. On the other hand, if one
considers a rotational band built on the same unique parity low K orbit in the
neighboring odd neutron nucleus, the alignment of two neutrons in that orbit
will be blocked by the prior occupation by the last odd neutron. There should
be no backbend, or at least a greatly reduced effect. If this is observed, one
deduces that the backbend in the even-even nucleus results from neutron
alignment. This technique has been exploited in many nuclei and an example
is given in Fig. 8.18. Here we see a case where backbending occurs in the even
nuclei 164Yb. In the neighboring odd nucleus 165Yb, the band based on the h9/2
orbit also backbends, but not that based on the im orbit. Thus one deduces
that the backbending in 164Yb nuclei is caused by the i13/z neutrons and not h9/2
neutrons, at least at this transition frequency.
The study of high-spin states, backbending, and related phenomena in
heavy nuclei has become a major area of activity in the last decade and an
enormous literature has developed. It is not our purpose here to summarize
either it or the many refinements and improvements that have been made to
this simple picture. Our aim is rather to provide some insight into the
"traditional" background. Suffice it to say that much more sophisticated
calculations are now standard, involving a "cranked" or rotated shell model
that quantitatively accounts for the transformation from the body-fixed Nilsson
scheme to the actual rotating system in the laboratory. Some of these calcula-
tions include the effects of axial asymmetry or hexadecapole deformations as
well.
Empirically, double and even triple backbending has also been observed at
higher transition frequencies. This has been interpreted in terms of the
successive alignment of first one pair of particles along the rotation axis and
then a second pair and sometimes even a third. Note that this reduces the
overall pairing strength. At very high rotation, the alignment of all pairs can
lead to a situation of complete pairing collapse. The study of unpaired intrinsic
states is a currently active field.
Another phenomenon which has recently attracted much attention is that
of superdeformation. This refers to a state at much larger than normal
quadrupole distortion, typically ft ~ 0.6. It has now been widely observed at
spins J ~ 40-60 in many rare earth nuclei and, possibly, at lower spins in the Hg
region. Its explanation involves steeply downsloping Nilsson orbits, stemming
from the second oscillator shell above the valence one, that descend, with
increasing ft, into the low-lying spectrum. Their occupation tends to stabilize
very large deformations.
Clearly, as ever higher spins are sought and attained, increasing sophistica-
tion in data acquisition and analysis is required and a remarkable richness of
phenomena unfolds. From all this has developed a deeper understanding of
rotational particle coupling, nuclear shapes and potential energy surfaces, the
phenomena of superdeformation and pairing collapse, and so on. The reader
is referred to a fascinating literature.
 9
MICROSCOPIC TREATMENT
OF COLLECTIVE VIBRATIONS

Although there are many microscopic approaches to deriving the structure,

energies, and systematics of collective states from the shell model, there is one
that has been widely used for nearly three decades, is still commonly encoun-
tered, and is easily adaptable to higher-order treatments. Moreover, it clearly
shows the basic microscopic physics that must be at the heart of any approach.
The technique referred to appears in two forms, the Tamm-Dancoff'Approxi-
mation (TDA) and the Random Phase Approximation (RPA). Doubtless, the
reader who is at all versed in nuclear structure physics has encountered
calculations carried out in the RPA or references to such techniques. Except
to the practitioners of this art unfortunately, the phrases TDA and RPA tend
to elicit fear and mystery. Those sections of theoretical papers describing such
calculations are frequently glossed over, and the reader quickly leaps to tables
or figures of the results. This is unfortunate for two reasons. First, these
sections often contain important information on the input physics (e.g., single-
particle energies, interaction strengths, etc.). Second, these techniques are
actually rather easy to understand if they are presented in a simplified form
that illustrates the essential physics rather than in an abstract formalism
designed to cover every generalization. In the next few pages we shall present
a simple derivation and discussion of the basic ideas involved, and then
illustrate the techniques with a particular calculation for rare earth nuclei. As
the reader will see, the end result will not only be a set of predictions for
comparison with experiment but a deeper understanding of the microscopic
nature of many aspects of collective behavior as well as of a very useful, but
often obscure, tool. It will then be easy to make predictions of the basic
structure of particular collective states without detailed or complex calcula-
tions. Simply by visual inspection of a Nilsson diagram, we will be able to
predict the energy behavior of collective vibrations (e.g., 7 or octupole), the
systematics of their collective properties (e.g., B(E2) or B(E3) values), and
even such details as whether they should be seen in single nucleon transfer
reactions. In short, a simple understanding of the basic ideas behind the
microscopic structure of collective states will give us the ability to anticipate,
without calculation, many results of detailed RPA or TDA calculations.
The following derivation, inspired by Lane, may seem rather formal and
abstract, but it is in fact easy to follow and leads to some very simple, powerful,
and useful results.
313
314 Collectivity, Phase Transitions, Deformation

9.1 Structure of Collective Vibrations

To start, we denote the ground state wave function of a many-body system by
*F0. *P0 is an eigenfunction of a Hamiltonian, H, such that

We call *P0 the vacuum state and will see exactly what this means later.
Consider now an arbitrary operator OJ that acts on *F0, giving a new wave
function ¥a

Now, suppose, and this is the key point on which all else depends, that Oaf
happens to satisfy the operator relation

Although we have yet to specify what form OJ must have for Eq. 9.3 to be
obeyed, or what coa is, let us center our attention for the moment on the
implications of this equation. Writing out Eq. 9.3 explicitly and acting on 4*,,,
we have

Using Eq. 9.2, this is

Using Eq. 9.1, we have

or, since E0 is a number, not an operator,

Using Eq. 9.2 again, we obtain

or, finally

Thus, 4*a, the result of acting on »F0 with OJ

a) is also an eigenfunction of H, and
b) has an excitation energy a>a with respect to the ground state Eg.
Thus, Oaf is a creation operator for the excitation a. This basic idea, plus some
simplifying assumptions about the structure of OJ, is the basis for the TDA
and RPA.
To proceed, we obviously need to find operators O^ that satisfy Eq. 9.3.
Alternatively, we can ask what the condition is such that Eq. 9.3 is satisfied. To
see this, let us expand OJ in another set of arbitrary operators that form a
complete set:
 Microscopic Treatment of Collective Vibrations 315

That is, the wave function for excitation a is

This is basically a definition of the expansion coefficients Xar. Using Eq. 9.5
in Eq. 9.3 gives

where, to avoid confusion, we use different subscripts in the independent

summations on the two sides. Since the X coefficients are just numbers, we
have

Now, we define a set of coefficients, Mrs, by the relation

and substitute this into Eq. 9.7:

or,

This, however, is just a matrix equation that must be satisfied for each r and
can be written

where M is the matrix whose elements are M and X is a column vector with
TS OC

elements Xm.
Thus, Eq. 9.4 and the condition on the OJ, Eq. 9.3, follow if Eq. 9.9 holds.
Alternatively, if Eq. 9.9 is true, then defining the OJ by Eq. 9.5, we find that
Eqs. 9.3 and 9.4 are obeyed.
At this point, it is probably not clear why we have indulged in this process of
piling definition upon definition. The aim was to produce Eq. 9.9. The reason,
and the practical use of all this is as follows.
The basic idea of the TDA and RPA (or higher-order approximations) is
a) to make assumptions about the operators O, defined in Eq. 9.5
b) to then use the definition given in Eq. 9.8 to solve for M (that is for the
elements Ma)
c) to then use these A/w's in Eq. 9.9 to solve for the XJs.
By Eq. 9.5 we then know the detailed structure of the operators Oaf that create
the excitation *Fa! If this seems amazing and magical, good. If it seems abstract
and artificial, be patient.
316 Collectivity, Phase Transitions, Deformation

To remove a little of the magic (but hopefully not the awe), let us consider
two simple assumptions for the O/s and see what results. Suppose the Or's
have the schematic form:

O, creates a particle-hole excitation by raising one nucleon to orbit;

Or creates and destroys a particle-hole excitation, denoted, for the moment,

simply by j. Later, when it becomes necessary to specify both orbits involved
in excitation ";'," we will expand the notation.
Approximation or assumption (i) is called the TDA and (ii) is called the
RPA. Before proceeding, we use the general forms of (i) and (ii) to understand
the meaning of the TDA and RPA.
We need only to recall that ¥„ is the ground state. That is, any destruction
operator acting on *¥„ must give zero:

For the TDA (assumption (i)), this is equivalent to

This will hold if A^ = 0 for all r. Therefore, this *F0 has no particle-hole
excitations r. If it had any, Ar could destroy one, giving a nonzero wave
function that did not contain a particle in that orbit. Thus, *P0 must be a closed
shell nucleus. Alternatively phrased, *P0 has no built in correlations. The TDA
therefore corresponds to creating excitations from a very simple, uncorre-
lated, uncollective ground state.
Assumption (ii), the RRA, on the other hand, means that we define OJ in
terms of Eq. 9.5 by

where, for convenience, we have separated those arbitrary operators "Or" that
are particle-hole creation operators (denoted A/) from those that are
particle-hole destruction operators (denoted Ar). Thus, we have

But by definition, *P0 is the ground state, implying that

for those r for which Ym * 0. But Arf creates a particle-hole excitation r.

Equation 9.14 holds then if *F0 already contains such a particle-hole excitation
so that Ar* is "blocked." Equation 9.13 can also be satisfied by cancellations of
several terms (Xm and Ym terms), which again implies the existence of
particle-hole excitations in VP0. Thus, although *P0 is itself the ground state
 Microscopic Treatment of Collective Vibrations 317

Fig. 9.1. Highly schematic illustration of the difference between the TDA and RPA approxima-
tions to the microscopic structure of collective excitations.

wave function, it already has some built in excitations (or correlations, or

collectivity). We illustrate the TDA and RPA in a very schematic way in Fig.
9.1. We expect that the RPA will be a somewhat more realistic or better
approximation, at least for nonclosed shell (or subshell) nuclei.
Now, let us use the approximations (i) and (ii) to see how they allow us to
solve for M, and hence for OJ, vPa, and a>a. If we do this, we will then know
both the microscopic structure of the excitation a directly in terms of the
occupation of the single-particle shell model orbits and the excitation energy
of this excitation!
For simplicity, we consider the TDA, approximation (i). Then, Eq. 9.8 is

Multiplying by A. on the left, writing out the commutator, and taking the
matrix element of Eq. 9.15 for state *F , gives

Recalling that A.^ and As simply create and destroy particles in orbit s, and
abbreviating wave functions *Fs = A^^ by s, (and similarly for /'), we have

where we also used the fact that the quantities Mrs are just numbers.
Now, the right side of Eq. 9.16 vanishes unless r = z (since f or r * i it contains
the factor (V0\ A .Arf 4*) = {4y0 A. *V), which vanishes by orthogonality since
HMias no particle in orbit;' to be destroyed). For r = /', we have on the right side,
318 Collectivity, Phase Transitions, Deformation

But the second term on the left, also

vanishes unless i = s, in which case it is given by the ground state energy E0.
Thus, we have the simple result

since we can choose to set E0 = 0. The particle-hole (or quasi-particle) energies

contained in H are energy differences between empty and filled orbits. We call
these energies Ae. In the RPA, a similar derivation yields a slightly more
complicated result that the reader can easily work out using analogous argu-
ments.
This result, Eq. 9.17, simply states that the matrix elements M.a of the matrix
M that, via Eq. 9.9, give the coefficients (vectors) Xa and the energy coa
defining the excitation a, are equal to matrix elements of the Hamiltonian
between states i and s. In order to calculate this explicitly, we need only to
make some choices for H = Ae + V. Starting from a shell model or Nilsson
model that gives the particle-hole energies Ae., we need only specify the
interaction V. Suppose we make the reasonable assumption that V is a
quadrupole interaction. Then,

where the Q's are proportional to the transition quadrupole matrix element
from the ground state to a particle-hole state with the particle in orbit i or j.
That is, Q. means (i IQ, ¥„> = (i \ r% | *P0). C is the strength of the interaction.
For an attractive interaction, C < 0. To keep the notation simple, recall that we
have specified each particle-hole excitation by a single subscript. Technically
this is incomplete; each such excitation involves elevating a particle to some
empty orbit and vacating a filled one. As long as no confusion results, we shall
keep to this notation but the reader should keep in mind that each Ae. involves
a pair of orbits and the energy difference between them.
Substituting Eq. 9.18 in Eq. 9.9 and using Eq. 9.17, we have

or

or

This must be satisfied for all r, so

or
 Microscopic Treatment of Collective Vibrations 319

or, finally, we get the expansion coefficients or amplitudes for the wave
function Ya of Eq. 9.6

TDA

To find the eigenvalues coa, we multiply Eq. 9.19 by Qr and sum over r,
obtaining

But 'LXar Qr on the left is identical to E^A^ Qs on the right, so cancelling

these, we have

TDA

which is the desired result. Note that the wave functions specified by the Xm
and the energy (oa are now written in terms of completely known quantities:
single-particle energies and single-particle transition quadrupole moments.
Given a single-particle model (shell, Nilsson, etc.), they are easily calculable.
Indeed, we shall see shortly that, using these results, we can for example
virtually "derive" y vibrational wave functions by inspection of the Nilsson
diagram without detailed calculations.
We emphasize that in Eq. 9.19, the quantity £<2,^a,is J ust a number, so that

Fig. 9.2. Illustration of the solution to Eq. 9.20TDA. The energies ^correspond to those points
where the vertical curves cross the "1/C" lines. The lowest (rightmost) solution is the collective
one (Ring, 1980).
320 Collectivity, Phase Transitions, Deformation

the coefficients of the vibrational wave functions for the excitation a are
simply proportional to QJ(&er - <»a). This fundamental result is of great
importance.
Recall that the Qr are the quadrupole transition matrix elements connecting
the ground state with a particular (>th) particle-hole excitation. Thus the
amplitude for particle-hole excitation r in the vibrational wave function is
proportional to the matrix element for a quadrupole transition to this
particle-hole excitation and is inversely proportional to the difference be-
tween the particle-hole energy, Aer, and the vibrational energy, a>a.
Particle-hole excitations that require energies close to ct)a and have large
quadrupole matrix elements are strongly favored.
The eigenvalue Eq. 9.20 has several possible solutions that are labeled a.
These correspond to those energies coa such that the right-hand side takes on
the value 1/C. These solutions are illustrated in Fig. 9.2. One solution always
appears significantly lowered in energy. This is the collective vibration.
It is easy to see this if we take a highly simplified example. Suppose there
are N identical particle-hole energies Ae and that the Qr values are also all
equal. Then the lowest solution has the wave function amplitudes

where a is just a constant. By normalization of this N-component wave

function, we have a = (I/ V/VjQ,). This means that the wave function consists of
a sum of equal amplitudes for all particle-hole excitations, with all amplitudes
in phase. This is exactly the formal definition of a coherent or correlated wave
function, which has many comparable amplitudes contributing with the same
sign. To see how this collectivity manifests itself, let us calculate an E2 (or
quadrupole) transition rate from the vibration a to the ground state:

Hence,

which is N times the single-particle B(E2) value. Since this exhausts the total
strength, it follows that no other solution has any strength. Of course, this is an
extreme simplification, but it does demonstrate what is meant by collectivity
and coherence and how they arise from this microscopic formalism.
Finally, the eigenvalue for the collective solution in this degenerate case is
given from Eq. 9.20 by setting all (<aa- Ae) constant, giving

r-i

The vibrational energy is lowered (recall that C < 0 for an attractive residual
interaction), relative to the (common) particle-hole energy by the large amount
 Microscopic Treatment of Collective Vibrations 321

NCQr2 This leads to an alternate viewpoint on the structure and origin of the
vibrational wave functions since they are identical to those we obtained in
Chapter 1 for the mixing of a set of degenerate levels. The lowest wave
function (see Eq. 1.13) was a sum of equal amplitudes for all basis states: this
state was lowered by an amount proportional to the number of admixed states
while all of the others were raised in energy.
Before closing this discussion and looking at applications of these ideas, it is
worth recapitulating the key ingredients that lead to the results obtained:
• We discussed a simple but abstract operator formalism leading to the
essential Eqs. 9.3, 9.4, 9.5, and 9.8. Equations 9.3 and 9.4 show the
properties of the excitation a whose structure we are interested in.
Equation 9.5 defines the creation operators OJ and Eq. 9.9, with the
definitions of Eq. 9.8, gives a means for solving for the coefficients Xm
that define the wave function *Fa of Eq. 9.6, and for the energies a>a.
• We made simple choices for the/orm of the various operators Oaf. These
correspond to the TDA (Oaf contains only particle-hole creation opera-
tors) and the RPA (Oaf contains both particle-hole creation and destruc-
tion operators).
• We made a simple choice for the residual interaction V. This step is often
called a schematic model. Doing this allowed us to calculate the A^'s,
hence the OJ's, and hence the structure and energy (fi>o) of the vibra-
tional state YB.
The TDA represents the gross approximation that the ground state is a
pure, undisturbed, closed shell nucleus. This is generally not realistic and the
use of the RPA to incorporate ground state correlations is more common. We
therefore present (without derivation) the eigenvalue results analogous to Eq.
9.20TDA, for the RPA:

or
RPA

This is similar to the TDA expression except for the more complicated
factor multiplying the Q\ Note that, in the degenerate case (all Aerequal) this
gives

or

or, if all the Qr are equal,

Using Eq. 9.20TDA and labeling the energies a>a as ft>aTDA or fi>aRPA, we have
322 Collectivity, Phase Transitions, Deformation

if all the Qr are equal. This shows that <oaRPA is always less than coaTDA. The RPA
leads to greater collectivity.
We close this section by noting that the discussion has been phrased in terms
of spherical single-particle-hole energies e; and single-particle quadrupole
moments (or other moments for other choices of V, such as octupole mo-
ments). However, as hinted at a couple of times, the single-particle energies
can be Nilsson energies equally well, and the A.\*s and A.'s can be creation and
destruction operators for Nilsson orbits. Then *¥a is the ground state of the
deformed nucleus. A related point concerns our choice for V in our example
of a schematic model in Eq. 9.18. It seems to ignore the short-range parts of the
residual (nonsingle particle) interaction such as the pairing interaction.
However, the whole formalism is identical if the pairing interaction is incorpo-
rated into the definition of the single-particle or Nilsson energies by defining
these as quasi-particle energies instead of single-particle energies. Then the
TDA or RPA can give the structure of vibrations in deformed nuclei in terms
of amplitudes for specific Nilsson quasi-particle excitations. We thus see how a
rather formidable looking formalism leads simply and elegantly to an easy way
of deducing the particle (or quasi-particle) composition of specific collective
vibrations. This is one answer to the apparent dichotomy that we noted earlier
between the independent particle picture of the nucleus and the existence of
collective excitations and correlations. The key element, of course, is in a
sense inserted a priori by defining the operators Oaf as linear combinations of
single-particle (or two-quasi-particle) operators and by simplifying the
interaction V. Ultimately, the method is tested by its agreement with experi-
ment. This test has been passed many times, making the result a useful,
elegant, and powerful approach to the microscopic structure of collective
vibrations.

9.2 Examples: Vibrations in Deformed Nuclei

The formalism just developed is ideally suited for understanding the micro-
scopic structure of vibrations in spherical nuclei. Indeed, calculations for
nuclei such as Pb or Sn were among its first applications. Since these nuclei are
singly magic, the simpler TDA is a reasonable approximation and is frequently
used. Pioneer applications of this formalism to such cases were made by
Kisslinger and Sorenson and are described in many texts. We will turn instead
to the application of these ideas to deformed nuclei.
Departing from singly magic nuclei, the p-n interaction rapidly induces
configuration mixing that obscures shell structure. The onset of deformation
affects the single-particle energies, as we have seen in the Nilsson scheme. So
for j3 = 0.3, any remnants of the 50, 82, or 126 shell gaps almost totally
 Microscopic Treatment of Collective Vibrations 323

disappear. Excitations involving orbits from the shells below or above the
valence shell must be incorporated in realistic calculations, and the RPA
becomes a necessary refinement.
Even higher-order forms of the operator Oaf are sometimes applied and are
often called higher RPA (HRPA) for obvious reasons. Other approaches such
as that of Kumar and Baranger or the more recent refinement, the dynamic
deformation theory, maybe used. The full sweep of many-body theory encom-
passes many varied and complementary approaches.
One of the most interesting, informative, and physically transparent appli-
cations of the RPA is to the /bands in deformed nuclei. The relevant
calculations were carried out by Bes and co-workers in the early 1960s and
remain the standard for the microscopic structure of these modes.
The basis states here are naturally the Nilsson orbits. Pairing must be
included via a BCS calculation, so that, for the ground state, there is a
distribution of occupation amplitudes over several orbits near the Fermi
energy /I. This can be seen quantitatively in a modification to Eq. 9.20RPA. For
this discussion, we must specifically label both orbits involved in the quasi-
particle excitation as we must specify occupation amplitudes for each. Replac-
ing the particle-hole energy Aer with Ei + E., where the £'s are quasi-particle
energies defined by

we have, for the energies (oa, the equation

where Qtj now means (i \ r'Y^ \f). The wave functions are now linear combina-
tions of two-quasi-particle excitations rather than single-particle-hole excita-
tions, but the physical idea is identical. Note that in Eq. 9.22, E. changes more
slowly than e when (e.- A) < A. Therefore, a wider range of e. values and orbits
can contribute. Also, the pairing factor favors pairs of quasi-particle excita-
tions on opposite sides of the Fermi surface; as an analogue to particle-hole
excitations, this is not surprising.
The matrix elements of r2Y2±2 that determine the important y-vibrational
amplitudes are easily deduced by writing r2Y2±2 in Cartesian coordinates as

This is a field (operator) that does not change nz (there is no effect in the z-
direction) but that changes A by ±2. (For a table giving the structure of all the
low multipole operators in Cartesian form, see Mottelson and Nilsson, 1959).
Another practical selection rule is that the sum nz + nf + ny = N should be
conserved; otherwise the matrix element would involve Nilsson wave func-
tions [Nnz\] and [N'n'^A'] with N' = N±2. Such states are far apart (about 10
MeV). Thus, the important components of the y-vibration will be two-quasi-
particle excitations involving a Nilsson orbit within about 2 MeV of the Fermi
324 Collectivity, Phase Transitions, Deformation

surface differing by AK = ±2, AA = ±2 (N, nz unchanged) from another near the

Fermi surface.
Since both K and A change by the same amount, the projection E of the
intrinsic nucleon angular momentum will not change (AE = 0). In considering
which orbits satisfy this rule, recall that the ground state consists of a corre-
lated wave function *P0 = Xfy + X^+..., each of whose components describes
two particles in the same Nilsson orbit but with opposite K values so that the
resultant K = 0. Since K = A ± E, there two nucleons also have equal and
opposite A values (±A). Hence, i2Y2±2 will also have matrix elements satisfying
the above rule that connect orbits such as [1/2 521] and 3/2[521], since the
A = ±1, K = ±1/2, and K = +3/2 components differ by AA and &K = 2, respec-
tively.
Let us now apply these ideas to a couple of examples. Consider 184W. The
ground states of mi85W are 1/2~[510] and 3/2-[512]. This approximately locates
the neutron Fermi energy in the Nilsson diagram (see Figs. 8.1 and 7.4). Then
we deduce by inspection that the important neutron components of the y-
vibration in 184W are two-quasi-particle states of the form
l/2~ [510] ®3/2- [512], l/2- [510] <S>5/2~ [512]
and

The analogous proton amplitudes can be similarly deduced. To see how well
this estimate works, we show the detailed results of RPA calculations in Table
9.1, from which it is evident that we have indeed identified the most important
two-quasi-particle components.

Table 9.1. Principle neutron two-quasi-particle amplitudes (xlOO) for the y-vihration in several
rare earth nuclei*
154 160 1M 170 174Yb !78Hf 184W 186W
Two-quasi-particle states Gd Dy Dy Er
ll/2-[505]®7/2-[514] — 7 7 — —
— — —
11/21505]® 7/21503] 11 12 15 25 26 21 11 8
9/21514]® 5/21523] 9 7 — — —
—
— —
9/21514]® 5/21512] 25 23 21 20 14 — — —
9/21505] ® 5/21503] — — — — 11 16
— —
7/21523]® 3/21521] 19 12 — — — — — —
7/21514]® 3/21512] — 17 35 48 34 26
— —
7/21503]® 3/21501] — — — — — 7 20 26
5/21523] ® 1/21541] 9 — — —
—
—
— —
5/21523] ® 1/21521] 20 36 51 29 14 8 — —
5/21512]® 3/21521] — — —
9 7 — — —
5/21512] ® 1/21510] — — 9 31 56 56 32 24
3/2~[532]® 1/21530] 24 8 — — —
—
— —
3/21521] ® 1/21541] 9 — — — —
— — —
3/2~[521] ® 1/21521] 27 45 46 25 13 8 — —
3/21512]® 1/21510] — — — 8 13 22 66 70

•Bfcs, 1965.The table only includes amplitudes from the N = 5 shell.

 Microscopic Treatment of Collective Vibrations 325

As a second case, consider 6g170Er102. Here the principle neutron two-

quasi-particle components of the y-vibration would be expected to be

Comparison with Table 9.1 again shows that these comprise most of the main
neutron components of the j-vibration in 170Er. Perusal of this table shows how
various two-quasi-particle amplitudes systematically grow and decay across
the region as the Fermi surface rises, and how different amplitudes are favored
by the energy denominator and pairing factors in Eqs. 9.20RPA and 9.22.
It is sometimes possible to test such predictions using single nucleon trans-
fer reactions. Since the 183W target ground state is the l/2-[510] orbit, the (d, p)
reaction can only populate two-quasi-particle states in 184W of the specific form
1/2"[510] ® v/-Nils- Moreover, they will contribute proportionally to IP-, the
emptiness of yNils in 183W. In 183W, therefore, the hole state 5/2-[512] component
should not contribute significantly to o(d, p), for the /-vibration, while neutron
transfer involving the 3/2~[512] orbit should. (Of course, in 182W the 7-vibration
must have similar structure. Hence, 5/2"[512] transfer will be important for the
183
W(d, t)182W cross section.) In any case, the (d, p) cross section to the various
spin states of the 7 rotational band in 184W will be determined primarily by the
C coefficients for the 3/2~[512] orbit.
The explicit expression for o(d, p) into an even-even nucleus is not quite as
simple as given in Chapter 8 for an odd-mass final nucleus, since each state of
spin / can be constructed by coupling a component (/j) of the 1/2~[510] orbit
with one;2 from the 3/2~[512] orbit. In general, several/^ pairs can lead to the
same/. Their relative contributions are given by Clebsch-Gordon coefficients.
The population of a two-quasi-particle state in an even-even nucleus in (d, p)
is thus given by a generalization of Eq. 8.1

where the 0( are DWBA cross sections (/ = ;'±l/2). It can easily happen that
several /values contribute to this expression for a given final state. In the case
of 184W, for example, the J* = 2+ state can arise by coupling the target;' = l/2~
state with; = 3/2" or; = 5/2' components of the 3/2~[512] wave function; these
proceed by / = 1 and / = 3 transfer, respectively. For a higher spin target such as
167
Er (7/2+), there are even more possibilities. A 4- level can be formed by
326 Collectivity, Phase Transitions, Deformation

coupling the 7/2+ ground state with;2 = 1/2, 3/2, 5/2, 7/2, 9/2,11/2 (the upper
limit here is provided not by angular momentum conservation but by the
available orbits in the N = 5 shell where ;max = 11/2). This involves six
contributions, with /=!,!, 3,3,5,5, respectively.
Thus, generally the "fingerprint patterns" in even-even nuclei will be
somewhat "washed out" and less orbit sensitive than in odd nuclei. Neverthe-
less, the (d, p) data for population of the y-band in 184W does have a pattern
very close to that predicted for the transfer of a 3/2-[512] neutron Nilsson orbit,
and the corresponding (d, t) data in 182W confirms the presence of the 5/2~[512]
orbit in the y-vibration there. Note that not only can the important orbits be
identified in this way, but their amplitudes in the /-vibration may be deduced
(approximately) by comparing, for example, the cross sections into the
even-even nucleus to those for transfer of the same orbit into the neighboring
odd A nucleus. In this way, even the detailed structure of RPA calculations are
confirmed, at least for a couple of important amplitudes. Other amplitudes,
such as 3/2-[512] ® 7/2-[514] cannot be directly tested.
In a few special cases it is possible to test for amplitudes involving a non-
target-ground state orbit. As an example, 185W has a ground state 3/2"[512] and
a low-lying 1/2~[510] excited state. If 185W were stable, permitting the relevant
(d, t) reaction to 184W, it would be possible to study the l/2-[510] ® 5/2~[512]
component in the 184W y-vibration, because the (3/2~) ground state of 185W has
a strong Coriolis admixture of the 3/2 1/2~[510] state. (The case starting with
183
W is not the same, even though here too the 1/2[510] and 3/2~[512] orbits are
admixed, since the l/2~ 183W ground state cannot contain a 3/2 3/2"[512] admix-
ture.) In the 185W(d, t)184W case, the y band would be primarily populated by a
coherent combination of the components 3/2~[512] ® 7/2[514] and 1/2~[510] ®
5/2[512].
To return to our discussion of the structure of collective excitations, we see
how an understanding of the basic formalism and philosophy of the RPA
allows us to understand and even anticipate the detailed microscopic structure
of excitations such as the y vibrations of deformed nuclei. A little further
thought reveals five additional basic features of these excitations.
First, the unique parity orbit is generally unimportant in the y vibration,
since this vibration has positive parity and therefore both quasi-particles have
to be in a unique parity orbit. But, from the sequence of asymptotic Nilsson
quantum numbers, we see that any pair with AK = 2 also has Anz = 2, and
therefore will not be connected by a r2Y2±2 operator.
Second, since the important orbits are those within a certain distance of the
Fermi surface, the y vibrational wave functions will not change radically from
one nucleus to the next or even over a small region. This is almost a
requirement of a collective mode. We can go even further and predict how the
structure will vary. From our discussion of the relevant (large) matrix ele-
ments and of the role of the energy denominator, it is clear that a given Nilsson
orbit will, for some mass A, be high above the Fermi surface and will contribute
little. As A (Noi Z really) increases, this Nilsson orbit will drop closer to the
Fermi surface and gain importance. Later, it will become a hole state and begin
 Microscopic Treatment of Collective Vibrations 327

decreasing in amplitude as it recedes further from the Fermi energy. Thus, for
most two-quasi-particle excitations, a plot of their contributions against NOT Z
(whichever is relevant) will be bell shaped. Inspection of Table 9.1 shows that,
for most orbits, this is precisely the observed behavior. Amplitudes for two-
quasi-particle states such as 9/2~[514] and 5/2^[512] containing an excitation
from the h^ orbit from the next lower major shell peak early in the deformed
region. The 5/2~[523] ® 1/2"[521] component increases into the deformed
region and attains its maximum amplitude near N = 9S (164Dy), but contributes
little for N > 104. Finally, the 7/2 -[514] ® 3/2~[512] combination is negligible
until rather late in the shell, but becomes strong near N = 108 and again drops
off in importance near jV = 112.
Third, we note that for any given nucleus, the number of significant compo-
nents is surprisingly low, typically three to five for neutrons and slightly fewer
for protons (since the protons are filling the lower shell with fewer orbits).
Thus the word "collective" must be taken in context. One should not imagine
50 to 100 nucleons or amplitudes significantly involved in this or other collec-
tive excitations.
Fourth, one should address the question of how the collectivity (as meas-
ured, for example, by large B(E2:7-> g) values) arises if only a few orbits
contribute strongly. This is especially so since "single-particle" B(E2) values
in odd-mass deformed nuclei (i.e., B(E2) values where one particle changes its
Nilsson orbit) are much smaller than in "shell model" nuclei. The former are
typically 10~3 to l(h* s.p.u., since the shell model strength is fragmented by
configuration mixing in deformed nuclei. Yet B(E2:y->g) values are typically
=10 s.p.u. This can only arise then as a specific effect of coherence: the
dominant contributions to these E2 matrix elements must add in phase.
This can be understood (at least qualitatively) by expressing the structure of
the vibration in an equivalent but, in a sense, inverted picture that we alluded
to briefly earlier. Instead of conceiving the vibration as built up of components
that arise by operating with r2Y2±2 on the orbits near the Fermi surface,
imagine a set of closely spaced two-quasi-particle states at some excitation
energy (e.g., 2 to 3 MeV) that mix by an appropriate interaction (that need not
be specified). Then, by the arguments concerning two- and multistate mixing
in Chapter 1, one level will be pushed down and its wave function will have the
most coherent admixture of amplitudes. More specifically, if one starts with
the idealized case of a set of degenerate levels with equal mixing matrix
elements, the lowest state after mixing is described by a thoroughly mixed
wave function (see Eq. 1.13) with all amplitudes identical and with the same
phase. For the still schematic case of nondegenerate but equally spaced initial
states that mix with equal matrix elements, a similar result is obtained. For the
more realistic case of more or less random initial spacings but roughly
comparable ^Y^ matrix elements (since otherwise the orbits in question
would not be important in the /vibration), the same qualitative feature
persists. Sample diagonalizations show this. In particular, the lowest state
always has its various wave function amplitudes in phase. Thus the B (E2:y-> g)
328 Collectivity, Phase Transitions, Deformation

values will have the maximum possible collectivity consistent with this set of
orbits and matrix elements.
This leads to the fifth feature—the energy systematics of the /vibration—
which we can again deduce qualitatively without explicit calculation. We need
only refer once again to the aforementioned mixing calculations and recall that
(taking for simplicity the case of N degenerate levels mixing with each other
with equal matrix elements) one state is pushed down well below all others and
that its energy lowering is (A7 - l)V, where V is the mixing matrix element.
Thus the y vibration will be lower when there are more contributing (mixing)
states. This occurs precisely when there is an abundance of Nilsson orbits near
the Fermi surface with identical nz values so that many large r 2 Y 2i2 matrix
elements with A« = 0, AA = ±2 contribute.
It was shown in our earlier discussion of the Nilsson model that at the
beginning of a shell, the steepest downsloping orbits have high nz (for prolate
nuclei). Since these orbit energies are so strongly correlated with the extent of
the wave function in the z direction, even neighboring orbits will have different
nz values. This makes it difficult to find nearby orbit pairs with AA = ±2 and
An = 0. Near the top of a shell, the large changes in orbit angles for different
K values means that there will be fewer orbits and that they will be further
apart on average. Thus, both cases only permit a few important y-vibrational
amplitudes. Just before midshell, however, one encounters a region where
orbits with various nz and A values congregate so that there will be more
contributing Anz = 0, AA = ±2 orbit pairs, hence greater collectivity and a lower
y-band energy. This qualitative behavior is sketched in Fig. 9.3 (which also
includes actual values calculated by Bes). These may be compared with the
observed systematics shown earlier in Fig. 2.17. The similarity of both ob-
served and calculated patterns to our simplified estimates is remarkable. (The
sharp drop empirically observed in E near A = 190 is beyond this approach
since it involves the onset of /-softness and large yms values even in the ground
state.)
It is worth noting that this systematics stems from the specific detailed
microscopic structure of the y vibration. It is not simply a handwaving
statement that vibrations will be lowest, and be most collective, at midshell
where there are the most valence nucleons. Few of the valence nucleons
actually participate. Moreover, other excitations, such as the j8 vibration (see
the following), have radically different systematics.
Similar arguments can be applied to other vibrational modes such as octu-
pole (or hexadecapole vibrations). For example, for K - 0~ octupole vibra-
tions, the relevant operator is Y3 0 and hence A«2 = ±3 AA = ±0 amplitudes are
critical. Being of negative parity, octupole vibrations need two orbits of
opposite parity. Thus, they must involve the unique parity orbit, and therefore,
there will not generally be as many available two-quasi-particle states that can
contribute. Octupole vibrations thus tend to be rather high-lying and not very
collective. They should lie lowest early in the deformed region where there are
a number of A nz= ± 3, AK = 0 combinations involving excitations from the high
/ normal parity orbits into the unique parity orbits.
 Microscopic Treatment of Collective Vibrations 329

Fig. 9.3. Qualitative estimate of the systematics of 7, /3, and octupole vibrations across ashell. The
second panel shows the actual calculations for a yvibration from Bes (1965).

For neutrons in the rare earth region, for example, amplitudes such as
l/2-[530] ® l/2+[660], 3/2-[521] ® 3/2+[651], and 5I2~[S12] ® 5/2+[642] can con-
tribute. As the shell fills, the low K unique parity orbits, whose participation
is essential, begin to fill in the ground state, effectively blocking the collectivity.
Thus K = 0, 3" excitations should rise in energy. Late in the shell, the normal
parity orbits are all low K, while the nearest unique parity ones are high K.
Hence K - 0~ octupole excitations are inhibited. Now, however, K = 3~
octupole vibrations, with amplitudes satisfying An = 0, AA = ±3 such as
ll/2+[615] ® 5/2-[512] or 9/2+[624] ® 3/2-[521], begin to increase in collectivity
and drop in energy. Thus one expects the qualitative systematics shown in Fig.
9.3, where the K sequence of successive octupole excitations in a given nucleus
is seen to change from K = 0-3 to K = 3 - 0 as a major shell fills. This
inversion of K ordering across a shell is verified in detailed calculations by
Neergaard and Vogel and confirmed by experimental systematics. It is inter-
esting that this inversion arises from the interplay of the quadrupole (Nilsson
330 Collectivity, Phase Transitions, Deformation

diagram structure) and octupole (Y3±Jf operators) modes. As a consequence,

the locus of the lowest J*= 3~ excitation (regardless of its K value) will have an
undulating character. First the K = 0~ excitation drops. Then it rises, and the
K = 3" excitation drops in energy.
It is instructive to apply similar arguments to the j8 vibration. Here one
encounters a different situation because of the K = 0 requirement, for which
the relevant operator is iaY20. this has large matrix elements only between
Nilsson orbits with AN = ±2, An = +2, AA = 0. However, any AN - 2 matrix
elements will naturally be small because of the large energy denominators. It
is not surprising then that /? vibrations are generally less collective than y
vibrations.
The collectivity that characterizes these states must arise from a completely
different origin. It occurs as a consequence of pairing correlations according
to an idea outlined in Chapter 1. The pairing interaction smooths out the
Fermi surface, leading to a gradual f alloff in occupation amplitudes for Nilsson
orbits around the Fermi energy. Thus, the ground state consists of a linear
combination of two-quasi-particle components, each with K = 0 and with the
two particles coupled antiparallel in the same Nilsson orbit. We then need to
evaluate the matrix elements of i^Y^ not between two-quasi-particle (Nilsson)
states but between two different linear combinations of pair wave functions.
For simplicity, consider a case of just two orbits, a and b, mixed by the pairing
interaction, and let the ground state and an excited K = 0 state be orthogonal
admixtures:

The calculation of (^ I ^Y^ | vgs) is just a special case of the general result
discussed in Chapter if We have now

The last term vanishes since Y2 0 cannot change the orbits of both particles.
The first two matrix elements are just the quadrupole moments of two par-
ticles in orbits b and a, respectively. Thus,

and this is small unless the orbits a and b have very different quadrupole
moments.
As we have discussed, the orientation of a Nilsson orbit (and hence its
quadrupole moment) is closely linked to its up- or downsloping character. For
prolate nuclei, downsloping orbits are equatorial and have prolate quadrupole
moments, while upsloping orbits are oblate. So ft vibrations should be rela-
tively collective and low-lying only in regions of the Nilsson diagram where
orbits with very different slopes lie near each other. Inspection of Fig. 7.4
 Microscopic Treatment of Collective Vibrations 331

shows that at the beginning and end of major shells, regions of strongly up- and
downsloping orbits approach each other from different shells. (The fact that
they have different /lvalues is inconsequential, since the allowed r^Y^ matrix
elements in Eqs. 9.27 and 9.28 are diagonal: they do not connect the two orbits
involved.) Near midshell, there are virtually no such orbit combinations.
Thus, P vibrations should be low in energy at the start and end of a shell and
should peak (and have the lowest collectivity) near midshell. This is illustrated
in Fig. 9.3 and can be compared to the empirical situation in Fig. 2.17. The
qualitative agreement is reasonable.
The brief discussion of the TDA and RPA formalism in this chapter, and its
application to some of the most important low-lying vibrations in medium and
heavy nuclei is meant only as an introduction to the field. In practical
calculations there are numerous subtleties (e.g., detailed choices of single-
particle energies) and sophistications (e.g., higher-order operators, O r , than in
Eqs. 9.11 and 9.12, or other related approaches involving self-consistent many-
body theory). Nevertheless, the underlying physics is always similar to that
outlined here. It is hoped that the present discussion will have removed some
of the mystery from such calculations and indicated how simple arguments,
based on the form (operator) for each type of vibration and the available
single-particle orbits, can lead to reasonable deductions of the principle wave
function components, their mass dependence, collectivity, and energy sys-
tematics.
This page intentionally left blank
 PART IV

EXPERIMENTAL TECHNIQUES
This page intentionally left blank
 10
A FEW SELECTED EXPERIMENTAL TECHNIQUES

The extraction of nuclear structure information from experimental observa-

tions has been implicit in much of the foregoing discussion. On numerous
occasions we used the energies of certain levels, transition rates, ratios of
energies, or ratios of transition rates to glean an idea of the underlying
structure.
In this chapter we will briefly outline the key ideas characterizing a few
experimental techniques that illustrate the rich variety of experimental tools
available to elucidate nuclear structure. In no sense will the discussion be
complete either for any specific technique or as a summary of all possible tools
available. The choice of technique to be described is largely arbitrary, dictated
in part by the information each provides, but especially by the author's experi-
ence.
We have already used single nucleon transfer reactions, such as (d, p) or
(d, t), to help gain a detailed understanding of the behavior of odd-mass
deformed nuclei. The use of stripping and pickup reactions to study near-
closed-shell spherical nuclei actually came first. It was in such work that the
concept of spectroscopic factor was first developed (the analogous quantity for
deformed nuclei is C2j or (7.C/2). We will not dwell further on these techniques,
but will turn instead to a few others to show a variety of empirical approaches
to nuclear structure.
The basic data needed to probe nuclear structure is extremely varied. They
include nuclear masses and binding energies, level energies and f values,
reaction cross sections, relative and absolute y-ray transition rates, electric and
magnetic moments (especially quadrupole moments), j3 decay rates, nuclear
radii, and so on. Here, we will illustrate the experimental side of nuclear
structure physics by brief, mostly qualitative, discussions of Coulomb
excitation/'complete" spectroscopy, with emphasis on average resonance cap-
ture (ARC), and heavy ion reactions, in particular (H.I., xnj). In each case the
discussion will be elementary, aimed at introducing the basic ideas and ap-
proaches, the type of physics they provide, and the advantages and limitations
of each technique. No attempt will be made to encompass the highly evolved
sophistication in their current use.

10.1 Coulomb Excitation

This technique has, since its introduction in the 1950s, provided perhaps the
most abundant source of detailed information on absolute y-ray transition
335
336 Experimental Techniques

probabilities, especially B(E2) values and nuclear quadrupole moments. It

has thereby provided perhaps the most solid evidence for the presence,
evolution, and variety of collective behavior exhibited by atomic nuclei. With-
out demeaning the important role of inelastic scattering techniques and direct
lifetime measurements, the most salient data demonstrating the collective
vibrational and rotational character of heavy nuclei certainly comes from
Coulomb excitation studies. In particular, the data for the plot of
B(E2:2+1 —> 0+j) values in Fig,. 2.16, which so vividly pointed to collective
effects away from closed shells, came predominantly from Coulomb excita-
tion, as do nearly all of the absolute E2 transition rates connecting y and
ground bands in heavy nuclei shown in Fig. 2.18.
The basic aim of Coulomb excitation is to provide a well understood means
of exciting atomic nuclei in a controlled manner— via a tractable, nonnuclear
interaction— so that nuclear matrix elements themselves can be extracted
from the data. As its name implies, Coulomb excitation involves the process of
nuclear excitation solely by the electromagnetic field between an incident
projectile and a target nucleus. By bombarding a target nucleus of atomic
number Z2 with a projectile of atomic number Zv at an energy sufficiently
below the repulsive Coulomb barrier between these two positively charged
objects, one ensures that the projectile does not enter the target nucleus, and
indeed, that its distance of closest approach is sufficiently far away that the
effects of the nuclear force are negligible. (We will briefly mention a variant of
this technique later where the latter condition is relaxed so as to exploit certain
quantum mechanical Coulomb-nuclear interference effects.) The electromag-
netic field is well understood and the process of Coulomb excitation can be
calculated explicitly. The principle computational tool for this is the so-called
deBoer-Winther computer code, developed in the late 1960s, which revolu-
tionized the modern use of the Coulomb excitation techniques by allowing a
simple and sufficiently accurate treatment of the complex multiple excitations
that occur when the incident projectile is itself a heavy ion.
The Coulomb excitation formalism has been lucidly described in great
detail in the classic paper by Alder, Bohr, Huus, Mottelson, and Winter. There
is no need to repeat their derivations and formalisms here. In fact, nearly all of
the key features, parameter dependencies, experimental arrangements and
constraints, possibilities, and limitations of the technique can be understood
and derived by a simple discussion of a few key ingredients. The main points
are illustrated in Fig. 10.1, which shows two possible trajectories for a projec-
tile (of energy Ep) incident in the vicinity of a typical target nucleus. The figure
defines a few basic quantities of importance in understanding the reaction. We
start by recalling that the Coulomb interaction is given by

where Z, and Z2 are the projectile and target atomic numbers, and r12 is the
time-dependent distance between target and projectile. Clearly, the probabil-
ity for exciting the nucleus by this Coulomb field is related to the magnitude of
V(r12) along the hyperbolic collision trajectory. The crucial ingredients are
 A Few Selected Experimental Techniques 337

Fig. 10.1. (Top) Diagram of some important parameters in the Coulomb excitation process.
(Bottom) Some E2 matrix elements involved in multiple Coulomb excitation.

then the product Z1Z2 and the sequence of values taken by r12. Since most of
the excitation will take place when the Coulomb field is strongest, the most
important physical parameter is p, commonly called the distance of closest
approach.
Clearly, the scattering angle 6 is closely connected with p: smaller values of
p during the collision lead to greater Coulomb interaction strengths and
therefore larger scattering angles. Conversely, the smallest distance of closest
approach, pmin, will be obtained when 9= 180°, and is given by

where m is the reduced mass of the projectile and target.

The essence of Coulomb excitation can then be summarized as follows: for
a given target nucleus, the excitation probabilities depend on £ , Z, , and 9. For
a given p, 9, they will be largest for projectiles of highest atomic number (Z1).
For a given projectile, they will be largest for the maximum energy Ep (and
therefore the minimum value of p), provided that Ep is well below the Cou-
lomb barrier. In practice, projectile energies up to approximately 80% of the
Coulomb barrier are used. Interestingly, Coulomb excitation has been of such
importance in nuclear physics that the desirability of exploiting this depend-
ence on Z1 and Ep in the form of higher-energy heavy-ion projectiles helped
338 Experimental Techniques

motivate the development of more powerful Tandem Van de Graaff accelera-

tors in the 1960s.
To further understand the Coulomb excitation process and the design of
Coulomb excitation experiments, let us consider the collision process in more
detail. A projectile passes near a target nucleus, and during the course of the
encounter, boosts this nucleus to an excited state. In the simplest such encoun-
ter, the projectile continues along its hyperbolic path and exits the region of the
target nucleus at an angle 6. The excited target nucleus subsequently decays,
nearly always by y-ray emission, to lower-lying levels.
A common simplification in the treatment of Coulomb excitation data,
known as the semiclassical approximation, assumes that the excitation of the
target nucleus has negligible effect on the hyperbolic orbit of the projectile. In
most cases, this is reasonable, since typical excitation energies produced in
Coulomb excitation are less than 1 MeV, whereas typical projectiles are
massive heavy ions such as 160,32S, and heavier species, with energies ranging
upwards of 50 MeV. The validity of this approximation is embodied in the
parameter 77 defined as:

The semiclassical approximation is valid if 77 » 1. This condition is

satisfied for most cases of practical intehhhhrhest. For example, for 60 MeV 16O ions
incident on 150Sm, 77 = 40.5.
Another important concept in the Coulomb excitation process is the excita-
tion energy AE imparted to the target. For a given projectile, Ep, 9, pmin, it is not
surprising that the cross section drops rapidly with increasing AE. This de-
pendence is embodied in the parameter <fj defined as the difference in incoming
and outgoing 77 values, 77., 7jr For small AE(AE/Ep« 1)

Since AE/E is typically = 10-2,£ typically ranges from 0 to 1.

The cross section for single step E2 Coulomb excitation can then be written

where df(9, £,) arises from an integral over the orbit of the projectile along the
classical trajectory. This expression embodies our previous comments. The
cross section increases as the distance of closest approach of the projectile
decreases, or as its charge increases. It also increases with increasing Ep : the
1/v factor is more than compensated by large df(6, £) values for small £ (large
Ep ) as we will discuss next. The df(8, £,) factor contains a sin-40/2 dependence
representing the underlying classical Rutherford scattering. (Often, Coulomb
excitation yields are given as a ratio to Rutherford scattering, removing this
strong dependence.) More importantly, df(9, %) decreases exponentially with
£. For large TJ, the integral of df(9, £) over 9 drops by a factor of 3 as \ goes
 A Few Selected Experimental Techniques 339

from 0 to 0.5 and by a factor of 10 as £goes from 0.5 to 1.0. In our example of
60 MeV 16O ions on 150Sm, these three <fj values correspond to excitation
energies of 0,1.5, and 3.0 MeV, respectively. Clearly, Coulomb excitation is a
tool for exciting low-lying states!
The preceding discussion implicitly assumed a single-step Coulomb excita-
tion process. There is no inherent reason, however, why multiple excitations
cannot occur during the collision process. Such multiple Coulomb excitation is
actually very common with heavy ion projectiles. The idea is illustrated in Fig.
10.1 (bottom). It is convenient to view this as a sequential process in which the
target is excited first to one level, then subsequently to another. However,
recognizing the extremely short collision times involved (= 10"22 seconds) the
uncertainty principle AtAE > h implies a large uncertainty in the system energy
during the collision. This permits the process of "virtual" excitation over short
time periods during which the conservation of energy is not required. Since
the probability of direct Coulomb excitation decreases exponentially with
excitation energy while virtual excitation decreases roughly linearly, it turns
out that the virtual excitation process is dominant in most multiple Coulomb
excitation situations. While high-lying states can be virtually excited during
the collisions, the range of final excitation energies produced is governed by an
exponential behavior. Nevertheless, the presence of virtual excitation both
offers an opportunity to excite levels not otherwise accessible and imposes a
complexity on the resulting analysis because the extraction of nuclear matrix
elements from Coulomb excitation data ultimately involves solving a set of
simultaneous equations that involve all matrix elements (and their generally
unknown relative signs) that connect all levels involved, virtually or otherwise,
in the excitation process.
The solution of these equations can be grea tly facilitated by certain experi-
mental tactics. One of these is based on the recognition that the only direction
fixed in space in a Coulomb excitation collision is that provided by the incident
beam direction. By conservation of momentum, those excitation processes in
which the projectile is scattered backwards at 180° can involve no transverse
momentum transfer to the target nucleus and, therefore, only the m = 0
magnetic substates of the final nuclear levels are excited. This immediately
reduces the number of simultaneous equations to be solved by a factor on the
order of (2/+1)2, where J is the average spin of the levels excited in the target
nucleus.
This quantum mechanical constraint can be imposed experimentally by
detecting the de-excitation y-rays in coincidence with back scattered projec-
tiles. Excitation of m * 0 substates increases slowly for 0 <180°, so that in
practice, one measures the backscattered particles at angle, ranging from 160°
to 175°. This technique uses annular detectors in which the incident beam
passes through a central hole and the backscattered particles are recorded in
the back angle annulus. Coincidence with backscattered particles has another
advantage: it preferentially selects those collisions with high excitation proba-
bilities (p small), thereby enhancing yields and leading to increased multiple
excitation so that B(E2) values for higher-lying states can be measured.
340 Experimental Techniques

There is another aspect of the fact that the incident beam defines a direction
in space. De-excitation y-rays have a particular angular distribution. In
principle, one should measure the yields at a series of angles and integrate over
angle. However, since virtually all Coulomb excitation proceeds by E2 excita-
tion (these are the strongest matrix elements connecting low-lying states), any
angular distribution involved can only include Legendre polynomials P((cos0)
of order 0, 2, and 4. Moreover, the coefficient of the quartic term is small in
most cases. Therefore, many Coulomb excitation experiments are performed
with the y-ray detectors situated at 55° relative to the beam axis since, at this
angle, P2 also vanishes, making the yields nearly identical to those obtained by
integrating over all angles.
Another simplification exploits the Z1 and Eprof dependence of Coulomb
excitation cross sections by carrying out a sequence of experiments starting
with light, low-energy projectiles. Typically, in even-even nuclei, Coulomb ex-
citation by protons or a-particles excites only the first level or the first couple
of levels. It is easy to extract the nuclear matrix elements for these. Indeed, it
was in the early days of Coulomb excitation, when only these projectiles were
available, that many B(E2:2+1 -» 0*1) values were first determined. With these
matrix elements in hand, a second experiment is performed, using a heavier,
and/or more energetic, projectile to excite the next set of levels. This process
continues until all of the matrix elements are deduced sequentially. Most
Coulomb excitation studies of heavy nuclei have been carried out in this
manner. In recent years, however, the development of sophisticated search
algorithms in conjunction with the deBoer-Winther code has led to the possi-
bility of simultaneously fitting an extensive set of Coulomb excitation meas-
urements made with high energy, heavy ion projectiles to extract a reasonably
unique and consistent set of matrix elements.
Thus far we have discussed experiments in which the presence of Coulomb
excitation is indicated by the observation of de-excitation y-rays. This is the
most common approach. It is hampered, however, by a built in ambiguity: one
does not generally know what fraction of each level's population (which is
deduced from the number of detected y-rays de-exciting that level) resulted
from direct excitations and from the decay of higher-lying levels. The detec-
tion of the outgoing, scattered projectile gives the number of excitations of
each level directly. Although the energy resolution of typical particle detec-
tors, or the poor efficiency of high resolution magnetic spectrometers, ham-
pers the use of this technique, it is often a valuable tool. It also leads to an
alternative to the preceding step-by-step approach to unravelling the E2 ma-
trix elements. In the traditional approach just described, this is achieved by
choosing the experimental conditions (Z1 and Ep) so that pmin attains different
values sequentially, the smallest corresponding to the most complex excitation
processes. From Fig. 10.1 it is clear that this can also be obtained with a single
projectile and fixed energy, simply by varying the angle 6 at which the scattered
projectile is detected. With the use of multiple particle detectors positioned at
different scattering angles, such a program can be carried out in a single
experiment, thus saving valuable running time and assuring identical experi-
mental conditions for each angle measured. The cost is that, for most angles,
the full set of magnetic substates must be included in the calculation.
Fig. 10.2. Data from two Coulomb excitation experiments. (Top) a-induced excitation of 160Gd
(Ronningen, 1977). (Bottom) Excitation of 232Th by 208Pb ions (Ower, 1982). The diagram shows
the experimental arrangement (see text). The spectrum at bottom illustrates the relative yields of
different y-rays while the boxes at the right give the population of two spin states as a function of
6. Note that the higher spin level is not seen at forward angles and is also much weaker at back
angles although, as expected, its yield increases faster with 9 than the J = 8 level.
342 Experimental Techniques

A final criterion that controls the choice of many of the projectiles used in
Coulomb excitation is that one generally wants to avoid projectile excitation
itself. Figure 10.1 is drawn in the laboratory system, but in the center of mass
system, it is impossible to distinguish projectile and target. One therefore
usually chooses projectiles that have very high-lying first excited states, gener-
ally magic or doubly magic nuclei such as a particles, 16O or M8Pb. Sometimes
it is more convenient to reverse the kinematic conditions and use the nucleus
conceptually considered to be the target as a projectile. This is useful when the
appropriate projectiles are not available, but it is also used, by choice, because
of the different scattering angles then obtained in the laboratory system. For
example, by using a Pb projectile on a Mg target, all yields will be in the
forward direction. In some cases such reversed kinematics can be advanta-
geous.
To obtain a concrete feeling for Coulomb excitation spectra, it is useful to
inspect the two examples shown in Fig. 10.2. The top one nicely illustrates low
energy, a induced, Coulomb excitation with detection of the inelastically
scattered projectiles. The strongest peak corresponds to elastic scattering.
The 2+ and 4+ members of the ground band are also strong. The excitation
probabilities fall off very rapidly as more excitation steps are required and the
6+, 8 + ,... states are not seen. Weaker, higher-lying peaks correspond to the
vibrational levels, 2+r and 3". The former is much stronger because there are
two excitation routes and because it involves E2 rather than E3 excitation.
Also note that Coulomb excitation is almost always limited to natural parity
states [it = (-1)' such as 0+, 2+, 4+, 3", but not 1+ or 3+] simply because of the
angular momentum conservation rules. This is rigorous in a single-step
process and nearly always characterizes multiple Coulomb excitation yields
also.
The lower part of Fig. 10.2 shows the experimental set up for 208pb induced
Coulomb excitation of 232Th. Such massive projectiles impart considerable
recoil to the target leading to Doppler shifts in the y-ray energies that depend
on the angle of recoil. The "Doppler" detector tags a specific angle. The
typical Ge(Li) y-ray spectrum shown demonstrates the ability to populate high
ground band and even side band states (n=-, here). (The 2+ -» 0+ and 4+ -» 2+
y-ray transitions are actually the strongest in this reaction; their weakness in
the spectrum is due to the fact they are highly converted.) Figure 10.2 includes
typical yields as a function of the angle of the scattered projectile. Note the
strong increase with 9 corresponding to smaller impact parameters p.
To understand multiple excitation more thoroughly, let us consider the
excitation of the 2+ band in Fig. 10.1. Even for relatively light projectiles there
are two possible excitation routes, a direct one from the ground state and a
double excitation via the 2\ state. In typical experimental conditions in which
backscattered particles are selected, the probability of exciting the 2+t state
may well be greater than 50 percent per detected backscattered particle. In
rotational nuclei, the Coulomb excitation matrix elements, <2+r E212 + t ) and
<2+rl E210 + j>, are simply related by Clebsch-Gordon coefficients. The latter is
slightly larger. However, it also involves a larger A£ than the step linking the
 A Few Selected Experimental Techniques 343

2\ and 2+rlevels. Therefore, in typical deformed nuclei, these two excitation

routes, 0 + j -> 2+, -» 2+r and 0+, -» 2+r, may well be comparable. Since they
interfere coherently, and since that interference may be either constructive or
destructive, depending on the relative signs of the matrix elements, the net
yield of the 2+r level can vary from almost vanishing to several times the yield
obtained for the direct route alone. The measurement of the excitation
probabilities for this level can then directly yield even the signs of these
nuclear matrix elements. This is especially true if the measurements are done
at several different beam energies, since the individual excitation routes have
a different dependence on beam energy. The signs of such matrix elements can
provide critical tests of certain nuclear models.
A second application of the multiple excitation process involves the recog-
nition that one kind of two-step excitation involves not an excitation from one
level to another, but a change from one magnetic substate of a single level to
another magnetic substate. This is also depicted in Fig. 10.1 for a couple of
states. Recalling that each step in a Coulomb excitation process involves a
matrix element connecting the initial and final levels, it is clear that this step
will involve the quadrupole moment, (2+l \ E2 \ 2+j>, of the level. If one meas-
ures the B(E2:0+j -> 2+j) value sufficiently accurately (typically within one
percent), it is possible to extract the 2+4 quadrupole moment, including its sign.
In a few cases this technique, known as the reorientation effect, has even been
applied to the measurement of quadrupole moments of higher excited states
such as the 2+r level.
There are other ways of obtaining the same information, such as the use of
Coulomb-nuclear interference effects, which we will discuss in a moment, as
well as muonic atom studies. While there has not always been agreement
between the quadrupole moments extracted from these different approaches,
the use of Coulomb excitation remains a standard technique.
Some nuclei are now known to display not only quadrupole but hexadeca-
pole deformations—/J4 or eA shape components—as well. These were dis-
cussed in connection with the Coriolis mixing of unique parity levels in odd
mass nuclei in Chapter 8. Typical hexadecapole shapes were shown in Fig.
6.10. Such shape components are also of interest in evaluating the role of g
bosons in the IBA (see Chapter 6). Once again, if the E2 matrix elements are
all well determined from previous Coulomb excitation experiments, the dis-
crepancies in the yields of certain states, for example the 4+ level, can be
attributed to direct E4 excitation and the relevant E4 transition moments
determined. It is largely from such data that the systematics (Fig. 8.10) of
hexadecapole deformations throughout deformed nuclei, both in the rare
earth and actinide regions, has been mapped out.
Occasionally it is useful to bridge the gap between Coulomb excitation and
nuclear inelastic scattering by using projectile energies close to the Coulomb
barrier. In this technique, the scattering cross sections contain both Coulomb
and nuclear amplitudes, and more importantly, an interference term between
them. It has been shown that this interference term is very sensitive to the
magnitude of the quadrupole moment of, for example, the 2+l state. This is
344 Experimental Techniques

Fig. 10.3. Coulomb/nuclear interference effects in ascattering near the Coulomb barrier (Bemis,
1973).

illustrated in Fig. 10.3, which shows the excitation cross sections for the 2*l and
4 + j levels divided by the pure Coulomb excitation cross sections. Deviations
from unity represent the effect of the Coulomb/nuclear interference. This
technique has been applied to many nuclei, in particular those in the W region.
The measured yields can also give hexadecapole shape components because of
the large interference effects for 4+ levels.
Until recently, Coulomb excitation experiments were carried out almost
exclusively below or at the Coulomb barrier. The reasoning was that interfer-
ence with direct or compound nuclear reactions would obscure the simplicity
of the technique. However, the increase of Coulomb excitation amplitudes
with projectile energy is not vitiated by competition with other processes. And
higher incident energies can provide access to higher-lying excitations. This
point has recently be exploited in very high-energy Coulomb excitation studies
of collective states, such as giant resonances (single or multiple) where other
excitation amplitudes are small. A technique thought by some to have already
witnessed its greatest popularity is thus being rejuvenated.

10.2 Spcctroscopically Complete Techniques

The (n, y) reaction has been used for several decades to elucidate nuclear
structure. It has some severe limitations and some powerful advantages. Its
primary attribute is nonselectivity, which makes it complementary to selective
processes such as Coulomb excitation or transfer reactions. In certain forms,
such as average resonance capture (ARC), the (n, y) reaction is actually
designed to be completely nonselective and to provide an a priori guarantee of
disclosing all states in certain spins and excitation energy ranges. The basic
idea of the (n, 7) reaction can be seen in Fig. 10.4 (left). An incident low-energy
Fig. 10.4. (Left) Schematic illustration of the difference between thermal neutron capture and ARC. (Right) Examples of the spin values and relative averaged in-
tensities available in the ARC process on even- and odd-mass targets.
346 Experimental Techniques

neutron impinges on a target nucleus (of mass A) and is captured, forming a

compound nucleus of mass A+l at an excitation energy (neglecting recoil)
Ec = S(n) + En where S(n) is the neutron separation energy. This rather high-
lying, low-spin compound nuclear capture state is extremely complex and its
decay, governed by that complexity, is largely statistical. The capture state de-
excites either by relatively high-energy "primary" transitions that directly feed
low-lying states or by a myriad of different cascade routes passing through the
high density of levels that occur between » 2 MeV and the neutron binding
energy. In either case, low-lying levels are finally populated: they decay by
discrete "secondary" transitions. With few exceptions, the distribution of
primary transition intensities can be described statistically in terms of a so-
called Porter-Thomas distribution with one degree of freedom. This is a broad
distribution, characterized by a most probable intensity of zero. Thus, while
each primary transition immediately discloses a specific low-lying level, little
or no nuclear structure information can be deduced, and moreover, many
primary transitions will be too weak to be observed. Owing to the multipole
decomposition of the radiation field, the dominant primary transitions are
nearly always El in nature. Ml transitions are nearly an order of magnitude
weaker and other multipolarities are almost never observed. Therefore,
primary transitions can be useful in immediately giving certain spin parity
limitations.
Most neutron capture experiments are carried out at thermal neutron
energy (£n= 0.025 eV), where s wave capture is totally dominant. For most
nuclei, S(n) = 5 to 8 MeV, and the level density is extraordinarily high so that
level spacings of capture states are typically in the few e V range. If one of these
levels happens to coincide with Ec, the thermal capture cross section will be
very large. Usually, however, it will be much smaller, arising from a sum over
the tails of several nearby levels. This difficulty can sometimes be avoided by
"tuning" the neutron energy (e.g., by Bragg diffraction) to a "resonance" so
that Ec coincides with a level of appropriate spin. Since level energies at
capture state energies are statistically distributed, such a process usually
selects a particular isotope as well, thus eliminating the need for enriched
targets. On the other hand, such epithermal neutron fluxes are normally
orders of magnitude weaker than at thermal energies.
The obvious disadvantage of either thermal or resonant energy neutron
capture is the enormously wide distribution of primary transition intensities
and the fact that, in general, no nuclear structure information is obtainable
from them other than pinpointing the existence and energies of specific levels
and giving certain limitations on their spins. A technique developed in the
early 1970s and utilized frequently in the last decade is an extension of
resonance neutron capture that avoids many of its disadvantages while having
the unique property of providing the possibility of an a priori guarantee of
observing all levels in certain spin and excitation energy ranges. This is the so-
called average resonance capture (ARC) process. The idea is extremely
simple, as is illustrated in Fig. 10.4 (left). Suppose one carries out a sequence
of resonance capture experiments with steadily increasing neutron energies,
 A Few Selected Experimental Techniques 347

sequentially populating, say, 50 to 100 individual resonances, and then sums all
the spectra so obtained. The statistical Porter Thomas distributions of each
resonance would then be nearly completely averaged out and all states of a
given final spin would be populated with nearly the same summed intensity.
Such an approach, although conceptually appealing, is clearly tedious and time
consuming. An alternative that achieves the same end automatically is to use
a beam of neutrons that has been rendered specifically non-monoenergetic.
This can be achieved by passing a reactor neutron beam through certain filter
materials, notably scandium and ^Fe, which have minima in their neutron
scattering cross sections near 2 keV and 24 keV, respectively. Having passed
through such materials, a reactor neutron beam will have an energy centered
around 2 or 24 keV with a FWHM (typically 1 to 2 keV) determined by the
thickness of the filter material. The FWHM must be sufficiently broad so that
a large number (Nr) of resonances are simultaneously populated, but not so
broad that it degrades the energy resolution of the primary y-ray transitions
following capture. The width of these lines is a convolution of the detector
(usually Ge) resolution (typically = 5 to 6 keV at 6 MeV) with the energy width
of the incident neutron beam.
To illustrate the process more concretely, consider a typical deformed
even-even target nucleus. Since the ground state is 0+ and 2 keV neutrons are
predominantly s wave, the capture state will be J = 112*. For nuclei such as
155
Sm, 167Er, 183W, the level spacing of l/2+ states at 1 MeV is «10 eV. Therefore,
a 2 keV neutron beam with 1 keV FWHM will populate =100 resonances. The
fluctuations in the combined Porter Thomas distribution go approximately as
21'i~N~r. Therefore, instead of orders of magnitude variations in primary
transition intensities, the fluctuations in the ARC spectrum intensities will be
± 2/VlOO = ± 20 %. (Actually, this variance refers to the reduced intensities,
defined as IJE5^ since it is necessary to remove a secular energy dependence.)
If we consider just El primary transitions (they dominate Ml transitions by a
factor of = 6), the final states will be 1/2-, 3/2- (Fig. 10.4, right).
When the target is odd mass, there are two possible capture state spins.
Each will decay by El (and weaker Ml, which we ignore) primary transitions.
In the simplest approximation, neglecting a spin dependence of the level
density near the neutron separation energy, both of these spins contribute
equally. Thus, as illustrated in Fig. 10.4 (right), the relative reduced intensities
will fall into two groups differing by a factor of = 2. An ARC spectrum was
shown for 196Pt in Fig. 5.8. The 195Pt target has J* = 1/2-, so that the (y-wave)
capture state is Jc = 0~, 1". Decay of the 1- levels by El transitions lead to 0+, 1+,
and 2+ states. The 0~ level only feeds 1+ levels. Final states with J* = 1+ should
thus be twice as intense as 0% 2+ levels. Of course, as we have seen in many
models, low-lying 1+ states are very rare. So, in practice, we expect rather
uniform intensities. The good averaging in Fig. 5.8 is immediately apparent,
especially when contrasted with typical intensity fluctuations of one to two
orders of magnitude in thermal capture experiments.
The preceding examples correspond to 2 keV ARC measurements where ,y
wave capture still dominates. At 24 ke V, both s and p wave components are
348 Experimental Techniques

Fig. 10.5. Ratio of reduced primary intensities in ARC at 2 and 24 keV in 168Er, for states of known
parity (Davidson, 1981).

roughly comparable and a wider range of final J" values becomes accessible.
The ratio of reduced ARC intensities at 2 and 24 keV provides a nearly ideal
parity meter for the final states. An example is shown in Fig. 10.5.
The limitation on excitation energies accessible in ARC stems from two
sources. First, with increasing Ex, the level density increases, which may make
it impossible to resolve individual levels. At first this is not a serious problem
since an unresolved pair of levels appears in the spectrum as a single peak of
twice the area and is therefore readily identified. At higher excitation ener-
gies, however, and especially in odd-odd nuclei, the probability of missed
levels rapidly increases. Second, and more importantly, the secular decrease of
the observed reduced intensities / (recall, it is the quantity /r IE 5 that is
constant) eventually leads to a point at which they can no longer be safely
detected above the background.
It is important to realize the consequences of the ARC-based guarantee of
completeness. It is not simply a question of finding an additional level or two
beyond that obtainable with other techniques. Its most profound importance
lies not so much in finding all levels (of certain J" and Ex ranges), but in
knowing that one has found all levels. This rather subtle point is best discussed
by two illustrative examples. Consider the imaginary level scheme shown on
the top in Fig. 10.6 and suppose that the ARC process populates levels with
spins 2+, 3+, 4+ via El primary transitions. The figure clearly depicts a de-
formed rotational nucleus with an excited vibrational band containing at least
2+ and 4+ levels. However, the figure is ambiguous as to whether this is a 7 band
with an undetected 3+ state or a p" band with undetected 0+ state. Even though
0+ states are not accessible to the ARC process in our example, this question
can be unambiguously decided. If the ARC data discloses a 3+ level between
the 2+ and 4+ states, then the band is clearly a 7 vibrational excitation. Equally
certain, if no such 3+ level is detected, there must be an undetected lower-lying
0+ level below the 2+ state. (Only K* - 0+ bands have a spin sequence 0', 2 + ,
 A Few Selected Experimental Techniques 349

Fig. 10.6. Examples of "completeness" arguments using ARC. (Bottom left) All n = -levels in
168
Er; (Bottomright) Arrangement into rotational bands. The numbers of bands of each lvalue
above - 1700 keV (right of dashed line) are discussed in the text.

4+,....) In this case, an ARC experiment identifies a particular vibrational

structure and points to an unobserved lower lying level simply by the nonob-
servation of another level. This process of obtaining information by nonobser-
vation is unique to completely nonselective processes.
Another common situation concerns levels with several existing y'possibili-
ties, say 112-, 3/2-, 512-, in an odd mass nucleus. ARC at 2 keV (j-wave) will
populate only (and all) 1/2" and 3/2" levels. Therefore, if the level in question
is undetected, its spin can unambiguously be assigned as 5/2". Again, one has
an example of spin assignment by nonobservation.
The completeness inherent in the ARC technique can also provide impor-
tant nuclear structure information, even though the process itself gives no
direct information on nuclear matrix elements. An example concerns the
higher-lying negative parity states in the well-deformed nucleus 168Er. Here,
the lowest negative parity states were already known and assigned to rota-
350 Experimental Techniques

Table 10.1. 168Er. Numbers of negative-parity states and deduced rotational band structures!
J*
r T 3- 4~ 5-
Number of observed states 2 1 4 5 6
Unique allocation of these
to rotational bands
one £*=0~ band 1 — 1 — 1
one K*=l~ band 1 1 1 1 1
two .£*= 3~ bands — — 2 2 2
two K*= 4~ bands — — — 2 2
tThe tabulation includes all negative-parity levels with spins from 1 to 5~ above 1780 keV not already allocated
to rotational bands with band-heads below 1570 keV.
'Davidson, 1981

tional hands. Between 1400 and 2000 keV, an ARC measurement disclosed
the set of negative parity states shown on the left in the lower part of Fig. 10.6.
The 167Er ground state is 7/2+, so the capture spins are 3+, 4+: therefore E1(M1)
primary transitions following s and p wave ARC populate all l~-5~ levels (up to
2 MeV) and distinguish the 2', 5~ group from 3", 4~ states from the factor of two
difference in reduced intensities. Combined with other data, a full set of
unambiguous J* assignments was made, but no structural information about
these states was known. However, given just the knowledge that this set of
negative parity states is complete, one can not only deduce the number of
rotational bands present, but their precise K values and the number of levels of
spin 0~, without having observed them. The idea is illustrated in Table 10.1.
The top gives the number of levels of each J. The key point is that the number
of 5~ levels is one greater than the number of 4- states. There are only two ways
to produce a 5~ state without an accompanying 4~level, either in a K* = 0~ band
with levels 1~, 3~, 5~, or in a K* = 5~ band with levels 5~, 6~, 7,.... It is easy to rule
out the latter. There are two 1~ levels and only one 2~ level. Again, the only
way this can occur is if there is a K* - 0^ band. Thus there cannot be a K* = 5~
band. Having established that there is one K = 0~ band, we have accounted for
one 1", one 3~, and one 5" state. Since there are two 1" levels and only one
K* = 0~ band, the other 1- state must be the bandhead of a K* = I' band. This
accounts for one each of the 1~, 2~, 3~, 4~, and 5~ levels, and thereby exhausts the
1~ and 2' levels but leaves two 3" levels unaccounted for. Each of these must be
K* = 3~ bandheads that account for two additional 4" and 5~ levels. There are
still two levels of each of these spins unaccounted for, and thus there must also
be two K* = 4' bands.
This now accounts for all of the states disclosed in the ARC measurement.
We have deduced the exact number of negative parity rotational bands and
their K values, and the fact that there is only one 0~ band and no K = 5 bands
below 2 MeV in 168Er. And we have deduced all this without assigning a single
state to a single rotational band, attempting to fit any of them into specific
rotational energy patterns, or inspecting their y-ray decay. The key was the
knowledge that the set of levels was complete. Some other experiment that
 A Few Selected Experimental Techniques 351

disclosed the same set of levels without providing this knowledge could not
have led to the same analysis.

Completeness with (a, xny) Reactions and

Extremely High Resolution /ray Spectroscopy
Other techniques besides ARC can provide near guarantees of completeness.
Examples that have been well documented are the (n, n') reaction, the (a, xny)
reaction, and the (n, y) reaction where one detects the low-energy, secondary
transitions. In all these cases the completeness is virtually assured, although
the method by which this occurs does not provide a specific guarantee. In each
of these reactions, the y-ray cascade from the initial formation state passes
through so many routes down to the lower-lying levels that the population of
the latter is effectively averaged out. Then, if one detects the secondary y-rays
connecting these states and fits these y-rays into a level scheme using the Ritz
combination principle, perhaps accompanied by information from y-ycoinci-
dence data, one is nearly assured of locating all levels. This has been demon-
strated in many important experiments. However, note that there is a concep-
tual difference between these cases and the ARC situation. In the latter, the
levels are populated directly and the measurement itself discloses their exis-
tence. In these other cases, the existence of specific low-lying levels must be
deduced indirectly by building a level scheme through Ritz combinations. It is
possible to miss a state in this process if, for example, it decays solely by one y-
ray to the ground state and is not observed to be fed by a discrete transition
from above. There would then be neither identifying coincidence relations nor
Ritz combinations. Nevertheless, such a case is rare, and these reactions have
been used to obtain essentially complete level schemes.
The uniformity of population is exemplified by the 123Te(a, ny)126Xe reac-
tion in Fig. 10.7, where the final level populations from side-feeding y-rays are
plotted against Ea for /flnal = 4 and 6. The smooth behavior and the clear
correlations with excitation energy demonstrates the unlikelihood of missing a
particular level. Moreover, the excitation functions show a difference in shape
for the two 7flml values. Such a dependence is systematic and allows one to
assign final state spins. A typical level scheme from (a, ny) is shown in Fig.
10.8. The richness is readily apparent, as is the wide range of spins accessible.
In comparison with (n, y) and the (Heavy ion, xny) reactions to be discussed,
(or, jeny) occupies an intermediate niche with a wider spin range than («, y) and
access to more states of a given spin than heavy ion reactions.
We have mentioned the study of secondary y-rays with the (n, y) reaction.
Such experiments encompass all the standard techniques of y-ray Spectroscopy
including the measurement of y-ray energies and intensities, coincidence
relations, and angular correlations, as well as the associated studies of conver-
sion electrons. Coincidence studies are very important in deducing nuclear
level schemes. They provide keys to the relative order and placement of
specific y-rays that complement those provided by energy sums alone. Con-
version electron studies are important in providing multipolarity information
on particular transitions. Angular correlation studies of successive cascade y-
352 Experimental Techniques

Fig. 10.7. Excitation functions for / = 4 and J = 6 states in the reaction 123Te(a, ny)126Xe (Lieberz
1989). The yields (7SF are for "side-feeding" transitions that exclude discrete feeding. Note the
nearly monotonic behavior against Ef (labeled for each line in keV) and the dependence of the
slopes on J.

rays also provide information on level spins, but they are seldom as unambigu-
ous as conversion electron measurements.
We will illustrate some of these techniques by focusing on the exceptional
capabilities of two instruments at the Institut Laue Langevin in Grenoble,
namely the electron magnetic spectrometer, BILL, and the y-ray spectrometer,
GAMS. Both are characterized by very high resolution and reasonable
efficiency. BILL achieves this by magnetic focussing, while GAMS measures
y-ray energies by diffracting the y-rays at precisely measured angles. Solid
state Ge y-ray detectors normally have resolutions ranging from = 800 eV at
100 keV to = 1.5 keV at = 800 keV. The corresponding numbers for the GAMS
spectrometers in normal operation are = 50 eV and = 400 eV. The corresponding
y-ray energy precision is typically 50 to 200 eV for Ge detection and 1 to 20 eV
for GAMS. A portion of a typical GAMS spectrum is shown in Fig. 10.9. The
primary impact of such resolution is not the higher level energy precision that,
in any case, far exceeds the predictive capabilities of modern nuclear models.
It is rather in the avoidance of accidental energy combinations ("Ritz combi-
nations ") that can lead to incorrect y-ray placements and, thereby, incorrectly
deduced structure information.
Consider an example of a typical deformed nucleus in which 250 y-rays are
Fig. 10.8. Typical level scheme from the (a, wf) reaction (positive parity levels only). (Lieberz, 1989).
Fig. 10.9. Examples of GAMS and BILL data. Note the remarkable resolution and energy precision of GAMS and the sensitivity to transition multipolarities
resulting from the resolution of L and M lines with the BILL spectrometer. From Krusche, 1989 (GAMS) and based on data discussed in Gelletly, 1981 (BILL).
 A Few Selected Experimental Techniques 355

observed with energies up to 1 MeV. If each y-ray has an energy precision of

± 0.1 keV, then crudely, this set of transitions "blocks out" a total energy
window of 50 keV or about 5 percent of the total. Roughly speaking, this is the
percentage of transitions that might be involved in incorrect Ritz combina-
tions. If a typical level has four de-excitation y-rays, then one in every two to
three levels will have a mistakenly placed y-ray either feeding or de-exciting it.
Clearly, this can lead to incorrect spin assignments, and certainly to incorrectly
assessed nuclear structure information. In contrast, with GAMS, the probabil-
ity of an accidental, incorrect Ritz combination would be 1/4 to 1/2 percent, or
in other words, perhaps one y-ray in the entire level scheme would be mis-
placed.
The advantages of GAMS are actually more important than this example
illustrates. Its resolution, though energy dependent, is particularly good for
the low-energy y-rays (below ~ 300 keV) that often provide the most crucial
nuclear structure information. They occur in competition with higher energy
transitions that are kinematically favored (transition rates are proportional to
B(E2) x E5 ). The observation of a low energy transition therefore nearly
always signals the presence of crucial collective matrix elements that over-
come the phase space hindrance.
One of the most powerful methods for assignment of electromagnetic
transition multipolarities is the use of conversion electron process. A level de-
exciting by electromagnetic radiation may occasionally knock out an atomic
electron rather than emit a y-ray. The probability of this is highly correlated
with the energy of the transition, its multipolarity, and the particular electron
shell involved. Conversion electron emission probabilities are normally ex-
pressed in terms of so-called conversion coefficients, which are the ratios of

Fig. 10.10. K conversion coefficients and L subshell ratios for El, E2, and Ml transitions.
Lifetimes for transitions of "single-particle" strength (including conversion) as a function of tran-
sition energy are also shown (right).
356 Experimental Techniques

the probability of emission of an electron from a given shell (K, L, M, N,...)

compared to that of the associated y-ray. Figure 10.10 (left) illustrates the K
conversion coefficients for two Z values and El, Ml, and E2 multipolarities.
When L shell electrons are emitted there are several possible transitions, the
so-called L subshell electron lines, each corresponding to a particular energy
difference between electronic orbits in the atom. The energies of these lines
depend on the particular element and grow rapidly with Z: in heavy nuclei near
A =150 binding energies are = 60 keV for the K shell and = 10 to 12 keV for the
L subshells. The separation of the Lj:L2:L3 lines is = 1 to 2 keV. Transition
multipolarities can be deduced from K conversion coefficients if the relative y
and electron intensity scales are known. However, the safest and most sensi-
tive method is the ratio of L subshell electron lines, which avoids all recourse
to y-ray intensities. Figure 10.10 indicates the very different El, Ml, and E2
L subshell patterns for a 200 keV transition in a nucleus with Z = 70. (The
figure also includes, for convenient reference, the single-particle lifetimes for
Ml, El, and E2 transitions for y-rays below = 5 MeV for Z = 55.) Unfortu-
nately, solid state electron detectors are incapable of resolving most L subshell
energy differences and can be used only for the measurement of K shell
electron intensities, or perhaps, the K to total L shell ratios. In contrast, the
BILL double focusing spectrometer can easily separate L subshell transitions,
and frequently, the M subshells as well. Furthermore, its efficiency allows one
to observe some N and O lines. A portion of a BILL spectrum is included in
Fig. 10.9.

10.3 Heavy Ion Compound Reactions and High Spin States

All of the techniques discussed so far deal essentially with low spin states. In
the last fifteen years, however, a large fraction of nuclear structure studies has
centered on high spin states, and an incredible richness of phenomena, barely
touched on in this hook, has been uncovered. We will give just the briefest
introduction to some of the many approaches to this field.
A projectile incident upon a target nucleus carries an orbital angular
momentum of 1 = r x p. The study of high-spin states, which necessarily
involves the transfer of large amounts of angular momentum to a target
nucleus, will clearly benefit from the use of heavy ion projectiles. An early
example of this approach was the pioneering (a, xny) studies of Lark and
Morinaga in the 1960s. They were able to populate the ground state rotational
bands of many even-even deformed nuclei up to the 6+ and 8+ levels. These
early studies provided important information on rotational structure, supply-
ing impetus, for example, to develop higher-order terms in the rotational
energy expansion (see Chapter 6). A natural outgrowth of these studies was
the use of true heavy ion projectiles; 12C, 16O at first, and subsequently, nearly
all nuclear species up to and including 208Pb and U. In the last couple of years,
evidence for discrete levels up to spins of 60ft and beyond have been detected
and have led to the discovery of the phenomenon known as superdeformation.
Figure 10.11 illustrates the basic population ideas behind a (H.I., xny) reac-
 A Few Selected Experimental Techniques 357

Fig. 10.11. Schematic idea of the flow of population and the de-excitation mechanisms in (heavy
ion,;t«y) reactions.

tion. An incident heavy ion impinges on and is captured by a target nucleus,

forming a compound nucleus. This state is formed at extremely high spin and
excitation energy. It is invariably particle unstable. The first step in the return
to a quiescent state is the emission of a number of evaporation neutrons. This
process is favored over y-ray emission both because it involves the strong
rather than the electromagnetic interaction and also because the more massive
neutrons provide a more efficient way of removing angular momentum for a
given energy emission. (Recall our discussion in Chapter 1 of the difficulties
inherent in high multipolarity y-ray emission.)
This process exaggerates one of the key features of (H.I., my) reactions—
they lead to extremely neutron deficient nuclei. They do so primarily because
most heavy ion projectiles (such as 16O or 32S) have nearly equal numbers of
protons and neutrons. When added to a heavy target nucleus, a compound
nucleus with a relative deficiency of neutrons is produced. This occurs
because the valley of stability in heavy nuclei proceeds, on an average, by
adding more neutrons than protons. The additional evaporation of several
neutrons from the initial compound nuclear stale increases the neutron defi-
ciency. Thus, (H.I.,jcny) reactions are useful both for studying high-spin stales
and gaining access to nuclei not otherwise observable.
Neutron emission, which proceeds along a path slanted toward the origin in
Fig. 10.11, terminates when the nucleus eventually reaches an excitation
energy that no longer permits particle emission. The only recourse is y-ray
358 Experimental Techniques

decay that can carry off considerable energy but relatively little angular
momentum. It therefore proceeds along nearly vertical de-excitation paths.
Given the high complexity of the compound state just prior to y-ray emission,
there are a myriad of possible routes and the decay is characterized by a nearly
statistical "rain cloud" pattern.
There have been many studies in recent years of the average properties of
the statistical continuum of y-rays including measurements of their average
multipolarities, average energies, and so on. We will not delve into this here.
The net result is that, finally, states relatively near the yrast levels are reached.
At this point the de-excitation path must closely follow the yrast line. The de-
excitation converts to the emission of a sequence of E2 y-ray transitions
feeding down any of several nearly parallel rotational bands in the yrasl

Fig. 10.12. Typical (Il.I.,^nr)-based level scheme (Riley, 1988).

 A Few Selected Experimental Techniques 359

region. Eventually, the de-excitation of these "side bands" terminates at or

near their bandheads and the intensity feeds into the yrast sequence itself.
From here on all the decay proceeds simply by stretched E2 transitions in the
ground state band.
Experimentally, the techniques for the study of (H.I., xny) reactions clearly
center around the detection and placement of the discrete y-ray lines that de-
excite the various high-spin levels, as well as the characterization of average
properties of the continuum lines. In a way, the situation is exactly comple-
mentary to the (n, 7) reaction. In heavy ion reactions, the discrete levels
observed are "cold," that is, among the lowest-lying states of those spins
energetically possible. The reaction discloses states over an extremely wide
spin range. In (n, 7), on the other hand, "hot" excited states are formed,
extending several MeV above the yrast sequence and encompassing a variety
of excitations, but always with spins within a few units of the target spin.
A typical level scheme observed in a (H.I., xny) reaction is shown in Fig.
10.12. One notes the separation of the levels into sequences of quasi-rota-
tional bands extending over extremely large ranges of spin. Even though these
bands may mix with one another, the decay, as we have noted, is preferentially
down the rotational band sequences rather than to high spin yrast states. We
saw in Chapter 6 why the decay remains within each band as long as the mixing
is not very spin dependent.
Except in the case of superdeformation, most of these rotational bands are
characterized by similar inertial parameters, h 2/27, and similar nuclear shapes.
Since they also involve similar spins, there will be many 7-ray transitions of
nearly comparable energies. This poses a serious experimental problem. It is
only with the development of multifold coincidence techniques in the last
years that this process has reached the stage of experimental sophistication
required to produce level schemes such as those in Fig. 10.12. The typical
experimental arrangement consists of a target surrounded by a large array of j-
ray detectors, usually Ge detectors flanked by anti-Compton shields designed
to enhance the photopeak efficiency and discriminate against Compton ef-
fects. Such arrays are commonly called "balls" and may involve upwards of 50
to 100 detectors, more in some proposed versions.
The procedure is to set up an electronic demand for two-, three-, or fourfold
coincidences as a prerequisite for triggering the recording of an event. This
allows one to select specific cascades and different spin regimes. The details
are well beyond the scope of this book, but an indication of the differences in
"fold" spectra is shown in Fig. 10.13. There is clearly a great advantage, in
terms of spectral "cleanliness," in arrays that are large enough to provide
sufficient solid angle to permit efficient multifold triggers.
We have already described some of the interesting results from heavy ion
reactions, namely the phenomenon of backbending and quasi-particle align-
ment in conjunction with our discussion of the Coriolis interaction. Recently,
a particularly intriguing phenomena has been discovered: superdeformation
in which rotational sequences are observed with moments of inertia indicating
deformation values of near ft ~ 0.5-0.6. This was first observed in 1986, and in
Fig. 10.13. Example of the advantages of multifold eoincidence gating. Data from the reaction
1M
Sn(29Si,^ny)1')8Gd. Each successive (lower) spectrum incorporates an additional coincidence
requirement (Deleplanque, 1988). (Top) "Singles"; (Middle) Single coincidence gate, (Bottom)
double gated. The lines remaining in this spectrum are mostly from the superdeformed hand.
 A Few Selected Experimental Techniques 361

the short time since, numerous other superdeformed sequences have been
found. At first, only high spin levels were associated with such structure.
Recently, in the Hg region, superdeformed states with / ~ 10 have been
assigned.
Initial interpretations of the superdeformed states involve the concept of
energy gaps in the Nilsson diagram at large deformations caused by the
presence of strongly downsloping Nilsson orbits from the N' = N+2 oscillator
shell. These deformed "shell gaps" act in a similar way to spherical shell gaps
and lend greater stability to those nuclei whose Fermi surface is near the gap.
The detailed characterization of superdeformed levels and the correct predic-
tion of their systematics offers a challenge to modern theories, both of high
spin states and of the behavior of single particle levels well above the Fermi
surface. Critical to such predictions has been the development of sophisti-
cated "cranked" shell model calculations that properly account for the effects
of extremely fast rotational motion, along with the incorporation of specific
corrections for shell structure not included by simply specifying the deforma-
tion parameters. The effects of the monopole proton-neutron interaction may
well be important in a detailed understanding of some of the latter effects.

10.4 Heavy Ion Transfer Reactions

Another experimental technique for nuclear structure studies involving heavy
ions that has been used in the last decade, albeit to a lesser extent, involves
heavy ion transfer reactions. The basic idea of the reaction is illustrated in Fig.
10.14. As opposed to the compound nuclear formation that is characteristic of
(H.L.jcny) reactions, heavy ion transfer reactions are essentially direct. Imag-
ine an incident heavy ion passing rather far from a potential target nucleus
(region I). There will be little interaction. The orbit of the projectile will be
only slightly changed and the probability for particle transfer will be very
small. If, on the other hand, the projectile nucleus passes very close to the
target nucleus (region III), it will likely be absorbed, forming a compound
nucleus, with miniscule probability that an ejectile similar to the incident
projectile will be emitted. The effective cross section for one or two nucleon
transfer will likewise be very small. Hence, there is an "impact parameter

Fig. 10.14. Illustration of three trajectories, and the associated angular distribution, for a heavy
ion transfer reaction.
Fig. 10.15. Spectra for three heavy ion single neutron transfer reactions leading to 14!lSm (Bond,
1983).
 A Few Selected Experimental Techniques 363

window," in which the transfer cross section maximizes, corresponding to a

"grazing" collision (region II). Thus the expected angular distribution is the
so-called "bell shaped curve" shown in Fig. 10.14.
Such one and two nucleon heavy ion transfer data have mostly been
exploited to study the reaction mechanism. But they can also be used in some
interesting ways to preferentially populate, and therefore identify, selected
high-spin states. This selectivity, especially when combined with coincidence
yray detection, overcomes the inherently poorer energy resolution of heavy
ion reactions. The idea is to exploit the possibility of high angular momentum
transfer and the differences in the orbits of the transferred nucleon for differ-
ent incident projectiles. Figure 10.15 shows some characteristic results.
As opposed to (H.I., xri) compound nuclear reactions in which the incident
projectile is absorbed along with all its angular momentum, heavy ion transfer
reactions only deposit large amounts of angular momentum if the incident and
outgoing energies and orbital angular momenta (I = r x p) are quite different.
The (13C, 12C) reaction, with a Q value near zero, is similar to (d, p), populating
low and medium spin states (lowest panel of Fig. 10.15). In contrast, Q values
for the (160,15O) and (12C,nC) reactions are highly negative (the projectiles are
tightly hound), so under appropriate kinematic conditions (E , 6^), large /
transfers are favored (top two panels).
There is more to it than this. In 12C, the last neutron occupies a p1/2 orbit
while it fills a p.^ orbit in 16O. This leads to very different selectivity for
J = 1-1/2 and 7 = 1+1/2 final states in the two reactions as we see in the figure:
in the middle, the 9/2' (h9/2,/ = /-l/2) state is comparable in strength to the
13/2+ (i1M,/ = / +1/2) and the 112- level (t,^j = / +1/2) is weaker than either. In
contrast, in the top panel, the / = / + 1/2 transfers (13/2+, 7/2+) are strongly
favored over the / = / - 1/2 (9/2-). This spin-orbit selectivity also leads to
interesting polarization phenomena. Studies of a number of nuclei (144Nd,
149
Sm, several Er isotopes, etc.) have exploited this technique.
In this short introduction to some of the experimental techniques of nuclear
structure physics we have, of necessity, been forced to be selective, giving only
a few illustrative examples of the rich possibilities that exist. Completely
untouched on, for example, are: /? decay and the study of the spectroscopy of
fission product nuclei, which in conjunction with powerful modern isotope
separators, provides access to hundreds of extremely neutron rich nuclei on
the opposite side of stability to that accessible in heavy ion reactions; inelastic
scattering (except as touched on briefly in the discussion or Coulomb/nuclear
interference effects); the study of nuclear radii and isotope and isomer shifts
with laser-based techniques; the study of magnetic moments using integral and
differential perturbed angular correlation techniques; and the use of photon
induced nuclear excitations to study both giant resonances and the recently
discovered isovector collective Ml scissors mode. The reader is simply re-
ferred to the literature to pursue these or other topics as his or her interests
dictate. Hopefully the structural background provided by this book will
facilitate such an adventure.
This page intentionally left blank
 REFERENCES

This list includes two categories of references. One consists of those specifically cited in
the text (for example, as the source for figures or tables). (The author apologizes for
the disproportionate and semi-obscene number of citations of his own work here. This
reflects both the fact that the topics covered in the book tend to be those that I am
interested in and may have worked on, and a degree of laziness in choosing illustrative
figures from my own papers rather than searching the literature for the better ones.)
The other category gives a few classic articles and other excellent works, often text-
books, review articles, or compilations, which are suggested for additional study. All of
these are recommended. Papers that the author has personally found to be continually
and particularly helpful, some indispensable reference works, and a few others are
marked with an asterisk. No attempt is made to provide a complete reference list for
ideas and data discussed in this book. This list is for the benefit of the reader, not for the
recognition of earlier accomplishments. That is the task of appropriate review articles.
References are cited in the text by first author only, with date of publication.

Alaga, G., Alder, K., Bohr, A., and Mottelson, B.R. (1955). Mat. Fys. Medd. Dan. Vid.
Selsk. 29, No. 9.
*Alder, A., Bohr, B., Huus, T, Mottelson, B.R., and Winter, A. (1956). Rev. Mod. Phys.
28, 432.
Alder, K. and Winther, A. (1966). Coulomb Excitation (Academic Press, New York).
Aprahamian, A., Brenner, D.S., Casten, R.F., Gill, R.L., and Piotrowski, A. (1987).
Phys. Rev. Lett. 59, 535.
Aprahamian, A., Brenner, D.S., Casten, R.F., Gill, R.L., Piotrowski, A., and Heyde, K.
(1984). Phys. Lett. 140B, 22.
Arenas Peris, G.E. and Federman, P. (1988). Phys. Rev. C38, 493.
Arima, A., Harvey, M., and Shimizu, K. (1969). Phys. Lett. 30B, 517.
*Arima, A. and Iachello, F. (1976). Ann. Phys. (N.Y.) 99, 253.
*Arima, A. and Iachello, F. (1978a). Ann. Phys. (N.Y.) 111, 201.
*Arima, A. and Iachello, F. (1979). Ann. Phys. (N.Y.) 123, 468.
Arima, A. and Iachello, F. (1978b). Phys. Rev. Lett. 40, 385.
Arima, A., Otsuka, T., Iachello, F., and Talmi, I. (1977). Phys. Lett. 66B, 205.
Arvieu, R. and Moszkowski, S. A. (1966). Phys. Rev. 145, 830.
*Baranger, M. (1960). Phys. Rev. 120, 957.
Barrett, B.R. and Kirson, M. W. (1970). Nucl. Phys. A148, 145.
Bemis, C. E., Jr., Stelson, P. H., McGowan, F. K., Milner, W.T., Ford, J.L.C., Jr.,
Robinson, R.L., and Tuttle, W. (1973). Phys. Rev. C8, 1934.
Bengtsson, T., Ragnarsson, I., and Aberg, S. (1988). Phys. Lett. 208B, 39.
*Bes, D.R., Federman, P., Maqueda, E., and Zuker, A. (1965). Nucl. Phys. 65, 1.
Bes, D.R. (1961). Mat. Fys. Medd. Dan. Vid. Selsk. 33, No. 2.
Bes, D.R. (1963). Nucl. Phys. 49, 544.
Bes, D.R. and Sorensen, R. A. (1969). Adv. in Nucl. Phys. 2, 129.
Bohle, D., Richter, A., Steffen, W., Dieperink, A.E.L., Lo Iudice, N., Palumbo, F., and
Scholten, O. (1984). Phys. Lett. 137B, 27.
365
366 References

*Bohr, A. and Mottelson (1969, 1975). Nuclear Structure. Vols. I and II, Benjamin,
New York.
Bohr, A. and Mottelson (1953). Mat. Fys. Medd. Dan. Vid. Selsk. 24, No. 16.
Bond, P.D. (1983). Comments Nucl. Part. Phys. 11, 231.
Braun-Munzinger, P. (1984). Editor, Nuclear Physics with Heavy Ions, Nucl. Sci. Res.
Conf. Series, Vol. 6, Harwood Academic, New York.
Brink, D.M., de Toledo Piza, R., and Kerman, A.K. (1965). Phys. Lett. 19, 413.
Bromley, D.A., Gove, H.E., Paul, E.B., Litherland, A.E., and Almquist, E. (1957).
Can. J. Phys. 35, 1042.
Brown, G.E. and Bosterli, M. (1959). Phys. Rev. Lett. 3, 472.
*Brussaard, P.J. and Glaudemans, P.W.M. (1977). Shell Model Applications in Nuclear
Spectroscopy, North Holland, Amsterdam.
Bucurescu, D. G., Cata, G., Cutoiu, D., Dragelescu, E., Ivascu, M., Zamfir, N. V, Gizon,
A., and Gizon, J. (1989). Phys. Lett. 229B, 321.
Bunker, M. and Reich, C. W. (1971). Rev. Mod. Phys. 43, 348.
*Castanos, O. and Draayer, J. P. (1989). Nucl. Phys. A491, 349.
*Casten, R. F. (1985a). Nucl. Phys. A443, 1.
Casten, R. F. (1986). Phys. Rev. C33, 1819.
Casten, R. F, Aprahamian, A., and Warner, D.D. (1984). Phys. Rev. C29, 356.
Casten, R. F., Brenner, D.S., and Haustein, P.E. (1987). Phys. Rev. Lett. 58, 658.
Casten, R. F., von Brentano, P., and Haque, A.M.I. (1985b). Phys. Rev. C31, 1991.
Casten, R. F. and Cizewski, J.A. (1978). Nucl. Phys. A309, 477.
Casten, R. F., Frank, A., Moshinsky, M., and Pittel, S. (1988). (Editors), Contemporary
Topics in Nuclear Structure Physics, World Scientific, Singapore.
Casten, R. F, Heyde, K., and Wolf, A. (1988b). Phys. Lett. 208B, 33.
Casten, R. F, Kleinheinz, P., Daly, P.J., and Elbek, B. (1972). Mat. Fys. Medd. Dan. Vid.
Selsk., 38, No. 13.
*Casten, R. F., Warner, D.D., Brenner, D.S., and Gill, R.L. (1981). Phys. Rev. Lett. 47,
1433.
Casten, R. F., Warner, D.D., and Aprahamian, A. (1983). Phys. Rev. C28, 894.
*Casten, R. F. and Warner, D.D. (1988a). Rev. Mod. Phys. 60, 389.
Casten, R. F., Warner, D.D., Stelts, M.L, and Davidson, W.F (1980). Phys. Rev. Lett. 45,
1077.
Cizewski, J. A., Casten, R.F., Smith, G.J., Stelts, M.L., Kane, W.R., Borner, H.G., and
Davidson, W.F. (1978). Phys. Rev. Lett. 40, 167.
Cizewski, J.A., Casten, R.F., Smith, G.J., MacPhail, M.R., Stelts, M.L., Kane, W.R.,
Borner, H.G., and Davidson, W.F. (1979). Nucl. Phys. A323, 349.
Cohen, B.L. and Price, R.E. (1961). Phys. Rev. 121, 1441.
Cottle, P. D., and Bromley, D. A. (1987). Phys. Rev. C35, 1891.
Davidson, W.F., Warner, D.D., Casten, R.F., Schreckenbach, K., Borner, H.G., Simic, J.,
Bogdanovic, M., Koicki, S., Gelletly, W., Orr, G.B., and Stelts, M.L. (1981). J.
Phys. G7, 455.
Davydov, A.S. and Filippov, G.F. (1958). Nucl. Phys. 8, 237.
Deleplanque, M.A. and Diamond, R.M. (1988). GAMMASPHERE Proposal.
de Shalit, A. and Goldhaber, M. (1953). Phys. Rev. C92, 1211.
*de Shalit, A. and Feshbach, H. (1974). Theoretical Nuclear Physics, Vol. I, Nuclear
Structure, Wiley, New York.
*de Shalit, A. and Talmi, I. (1963). Nuclear Shell Theory, Academic Press, New York.
*Dobaczewski, J., Nazarewicz, W, Skalski, J., and Werner, T. (1988). Phys. Rev. Lett.
60, 2254.
Eisberg, R. and Resnick, R. (1974). Quantum Physics, Wiley, New York.
Eisenberg, J.M. and Greiner, W. (1970). Nuclear Models, Vol. 1, Excitation Mechanisms
of the Nucleus, Vol. 2, and Microscopic Theory of the Nucleus, Vol. 3, North Hol-
land, Amsterdam.
Endt, P.M. and Van der Leun, C. (1978). Nucl. Phys. A310, 1.
 References 367

Elliott, J.P. (1958). Proc. Roy. Soc. A245,128 and 562.

Elliott, J.P. and Harvey, M. (1963). Proc. Roy. Soc. A272, 557.
Fallieros, S. and Ferrell, R. (1959). Phys. Rev. 116, 660.
*Federman, P. and Pittel, S. (1977, 1978). Phys. Lett. 69B, 385 and 77B, 29.
*Gallagher, C.J. (1964). In Selected Topics in Nuclear Spectroscopy (ed. BJ. Verhaar),
North Holland, Amsterdam.
Garrett, J.D. (1984). Nucl. Phys. A421, 313c.
Garrett, J.D., Nyberg, J., Yu, C. H., Espino, J. M., and Godfrey, MJ. (1988). In
Contemporary Topics in Nuclear Structure Physics, p. 699 (volume referenced
elsewhere in this listing).
Gelletly, W. and Schreckenbach, K. (1981). Private communication.
Ginocchio, J. N. and Kirson, M. W. (1980). Nucl. Phys. A380, 31.
Gneus, G. and Greiner, W. (1971). Nucl. Phys. A171, 449.
Gustafson, C, Lamm, I.L., Nilsson, B., and Nilsson, S.G. (1967). Ark. Fysik 36, 613.
Hager, R.S. and Seltzer, E.C. (1968). Nucl. Data A4, 1.
Haxel, O., Jensen, J.H.D., and Suess, H.E. (1950). Z. Phys. 128, 295.
Hecht, K. (1964). In Selected Topics in Nuclear Spectroscopy (ed. PJ. Verhaar), North
Holland, Amsterdam.
*Heyde, K. (1990). The Nuclear Shell Model. Springer Series in Nuclear and Particle
Physics, Springer-Verlag, Berlin.
Heyde, K. (1989), Int. J. Mod. Physics A4, 2063.
Heyde, K., Jolie, J., Moreau, J., Ryckebusch, J., Waroquier, M., Van Duppen, P., Huyse,
M, and Wood, J.L. (1987). Nucl. Phys. A466, 189.
Heyde, K. and Sau, J. (1986). Phys. Rev. C33, 1050.
*Heyde, K., Van Isacker, P., Casten, R.F., and Wood, J.L. (1985b). Phys. Lett. 155B, 303.
*Heyde, K., Van Isacker, P., Waroquier, M., Wood, J. L., and Meyer, R. A. (1983). Phys.
Rep. 102, 291.
Holmberg, P. and Lipas, P.O. (1968). Nucl. Phys. A117, 552.
*Hornyak, W.F. (1975). Nuclear Structure, Academic Press, New York.
Iachello, F. and Scholten, O. (1979). Phys. Rev. Lett. 43, 679.
Iachello, F. and Talmi, I. (1987). Rev. Mod. Phys. 59, 339.
Janssen, D.R., Jolos, R. V, and Donau, F. (1974). Nucl. Phys. A224, 93.
Kerman, A.K. (1959). In Nuclear Reactions, Vol. 1 (eds. P. M. Endt and M. Demuer),
North Holland, Amsterdam.
*Kisslinger, L.S. and Sorensen, R. A. (1960). Mat. Fys. Medd. Dan. Vid. Selsk 32, No. 9.
Kleinheinz, P., Casten, R.F., and Nilsson, B. (1973). Nucl. Phys. A203, 539.
Krusche, B., Borner, H., Takruri, E, Robinson, S., Michaelson, S., and MacMahon, D.
(1989). Private communication.
Kumar, K. and Baranger, M. (1968). Nucl. Phys. A122, 273.
Kumar, K. (1982). In Contemporary Research Topics in Nuclear Physics (eds. D.H.
Feng, M. Vallieres, M.W. Guidry, and L.L. Riedinger), Plenum, New York.
Kuo, T.T.S. and Brown, G.E. (1966). Nucl. Phys. 85, 40.
*Lane, A.M. (1964). Nuclear Theory, Benjamin New York.
Leander, G.A., Dudek, J., Nazarewicz, W., Nix, J.R., and Quentin, Ph. (1984). Phys.
Rev. C30, 416.
•Lederer, C.M. and Shirley, S. (1977). Table of Isotopes, 7th Edition, Wiley, New York.
Lieberz, W., Dewald, A., Wirowski, R., Zell, K. O., and von Brentano, P., to be
published.
Lipas, P.O. (1962). Nucl. Phys. 39,468.
Ma, W.C., Quader, M.A., Emling, H., Khoo, T.L., Ahmad, I., Daly, P.J., Dichter, B. K.,
Drigert, M., Garg, U., Grabowski, Z.M., Holzmann, R., Janssens, R.V. F.,
Piiparinen, M., Trzaska, W.H., Wang, T.F. (1988). Phys. Rev. Lett. 61, 46.
Mampe, W., Schreckenbach, K., Jeuch, P., Maier, B.P.K., Braumandl, F., Larysz, J., and
von Egidy.T., 1978. Nucl. Instr. and Meth. 154, 127.
Mayer, M. G. (1950). Phys. Rev. 78, 16.
368 References

*Molinari, A., Johnson, M. B., Bethe, H.A., and Alberico, W. M. (1975). Nucl. Phys.
A239, 45.
Mottelson, B.R. and Nilsson, S. G. (1959). Mat. Fys. Skr. Dan. Vid. Selsk., 1, No. 8.
Nagai, Y., Styczen, J., Piiparinen, M., Kleinheinz, P., Bazzacco, D., von Brentano, P.,
Zell, K. O., and Blomqvist, J. (1981). Phys. Rev. Lett. 57, 1259.
*Nathan, O., and Nilsson, S. G. (1965). In Alpha, Beta, and Gamma Ray Spectroscopy,
Vol. 1 (ed. K. Siegbahn), North Holland, Amsterdam.
Neergaard, K. and Vogel, P. (1970). Nucl. Phys. A145, 33.
*Nilsson, S. G. (1955). Dan. Mat. Fys. Medd. 29, No. 16.
Nilsson, S. G. and Prior, O. (1961). Mat. Fys. Medd. Dan. Vid. Selsk 32, No. 16.
Nilsson, S. G., Tsang, C. F., Sobiczewski, A., Szymanski, Z., Wycech, S., Gustafson, C.,
Lamm, I.L., Moller, P., and Nilsson, B. (1969). Nucl. Phys. A131, 1.
Ower, H., Elze.Th. W., Idzko, J., Stelzer, K., Grosse, E., Emling, H.,Fuchs, P., Schwalm,
D., Wollersheim, H. J., Kaffrell, N., and Trautmann, N. (1982). Nucl. Phys. A388,
421.
Paar, V. (1979). Nucl. Phys. A331, 16.
Preston, M.A. and Bhaduri, R. K. (1975). Structure of the Nucleus, Addison-Wesley,
Reading, MA.
Preston, M.A. and Kiang, D. (1963). Can. J. Phys. 41, 742.
*Raman, S., Malarkey, C.H., Milner, W.T., Nestor, C.W., Jr., and Stelson, P.H. (1987).
At. Data and Nucl. Data Tables, 36, 1.
Ratna Raju, R.D., Draayer, J.P., and Hecht, K. T. (1973). Nucl. Phys. A202, 433.
*Riedinger, L.L., Johnson, N.R., and Hamilton, J.H. (1969). Phys. Rev. 179, 1214.
Riedinger, L.L., Smith, G.J., Stelson, PH., Eichler, E., Hagemann, G.B.,Hensley, D.C.,
Johnson, N.R., Robinson, R.L., and Sayer, R. O. (1974). Phys. Rev. Lett. 33,
1346.
Riley, M.A., Simpson, J., Sharpey-Schafer, J.F., Cresswell, J.R., Cranmer-Gordon,
H.W., Forsyth, P.D., Howe, D., Nelson, A.H., Nolan, P.J., Smith, P.J., Ward, N.J.,
Lisle, J.C., Paul, E., and Walker, P.M. (1988). Nucl. Phys. A486, 456.
Ring, P. and Schuck, P. (1980). The Nuclear Many-Body Problem, Springer, New York.
Ronningen, R.M., Hamilton, J.H., Ramayya, A.V., Varnell, L., Garcia Bermudez, G.,
Lange, J., Lourens, W., Riedinger, L.L., Robinson, R.L., Stelson, P.H., and Ford,
J.L.C., Jr. (1977). Phys. Rev. C15, 1671.
*Sakai, M. (1984). At. Data and Nucl. Data Tables 31, 399.
Scharff-Goldhaber, G. and Weneser, J. (1955). Phys. Rev. 98, 212.
Scholten, O., Iachello, F., and Arima, A. ( 1978). Ann. Phys. ( N.Y.) 115, 325.
Sheline, R. K. (1960). Rev. Mod. Phys. 32, 1.
*Schiffer, J.P. (1971). Ann. Phys. (N.Y.) 66, 798.
*Schiffer, J.P. and True, W. W. (1976). Rev. Mod. Phys. 48, 191.
*Stephens, F.S. (1975). Rev. Mod. Phys. 47, 43.
Stephens, F.S. and Simon, R.S. (1972). Nucl. Phys. A183, 257.
Tabor, S.L., Cottle, P.D., Gross, C.J., Huttmeier, U.J., Moore, E. F., and Nazarewicz, W.
(1989). Phys. Rev. C39, 1359.
*Talmi, I. (1962). Rev. Mod. Phys. 34, 704.
Twin, P., Nyako, B.M., Nelson, A. H., Simpson, J., Bentley, M.A., Cranmer-Gordon, H.
W., Forsyth, P.D., Howe, D., Morhtar, A. R., Morrison, J.D., Sharpey-Schafer,
J.F., and Sletten, G. (1986), Phys. Rev. Lett. 57, 811.
Van Duppen, P., Coenen, E., Deneffe, K., Huyse, M., and Wood, J.L. (1985). Phys. Lett.
154B, 354.
Van Maldeghem, J., Heyde, K., and Sau, J. (1985). Phys. Rev. C32, 1067.
Warner, D.D., Casten, R.F., and Davidson, W.F. (1980). Phys. Rev. Lett. 45, 1761.
Warner, D.D. (1984). Phys. Rev. Lett. 52, 259.
Warner, D.D. and Casten, R.F. (1982). Phys. Rev. Lett. 48, 1385.
*Warner, D.D. and Casten, R.F. (1983). Phys. Rev. C28, 1798.
Waroquier, M. and Heyde, K. (1971). Nucl. Phys. A164, 113.
 References 369

*Wilets, L. and Jean, M. (1956). Phys. Rev. C102, 788.

*Wilkinson, D. (1983). Editor, Collective Bands in Nuclei, Vol. 9, Progress in Particle
and Nuclear Physics, Pergamon Press, New York.
Wirowski, R., Yan, J., Dewald, A., Gelberg, A., Lieberz, W., Schmittgen, K.P., von der
Werth, A., and von Brentano, P. (1988). Z. Phys. A239, 509.
Wolf, A. and Casten, R.F. (1987). Phys. Rev. C36, 851.
Wu, C.L., Feng, D.H., Chen, X.G., Chen, J.Q, and Guidry, M.W. (1986). Phys. Lett.
168B, 313.
Wyss, R., Granderath, A., Lieberz, W., Bengtsson, R., von Brentano, P., Dewald, A.,
Gelberg, A., Gizon, A., Gizon, J., Harissopulos, S., Johnson, A., Nazarewicz, W.,
Nyberg, J., and Schiffer, K. Nucl. Phys., to be puhllshed.
Zamfir, V., Casten, R.F., and von Brentano, P. (1989). Phys. Lett. 226B, 11.
Zhang, J.-Y., Casten, R.F., and Brenner, D.S. (1989). Phys. Lett. 227B, 1.
This page intentionally left blank
 INDEX

This listing gives the locations of principal discussions of various ideas and
topics. Occasionally, casual references to an idea or technique (e.g., paren-
thetical comments in the text) are not listed. Also, themes running throughout
the text, sometimes implicitly, such as the attractive nature of the nuclear
force, the Pauli principle, two-state mixing, the nuclear potential, collectivity,
and so on are only cited where they are explicitly the principal topic. Ideas,
nuclei, and so on that are the subject of figures are cited by the page numbers
where those figures are discussed rather than where the figures themselves
occur. Greek letters and words are alphabetized according to their latin
spelling and acronyms (e.g., RPA, GAMS, CFP's) are generally given accord-
ing to the way they are commonly thought of: thus, CFP's are listed under "co-
efficients..." while GAMS is given according to the letters themselves. Such
decisions are often quite arbitrary.

Alaga rules, 172-183, 212-214, 228-230 charge independence, 12, 15

algebraic models, 3, 197-235 charge symmetry, 8
(o, xy),351 closed shells, 31, 49, 58, 68
ARC, 346-351 CFP's, 111-115, 149-154
axial asymmetry, 168, 175-197, 221-234 coherence, 23-25
collective models, 3, 34, 141-331
backbending (see decoupled bands, collectivity, 23, 24, 130, 320-322, 327
rotation alignment), 233, 307-313 completeness (spectroscopic), 344,
bandmixing, 175-183, 228-234 348-351
bell-shaped angular distributions, configuration mixing, 141-145, 249-252
361-363 conversion coefficients, 355, 356
B(E2) values, 41, 42, 118, 121, 126, Coriolis coupling, 24, 175, 265, 266,
151-154, 159-161, 170-183, 279-312,326
189-193, 206, 207, 211, 212, 217, and backbending, 307-312
221-233, 244, 245, 327, 336 decoupling parameter, 282-284
0-decay, 29 hexadecapole deformation, 296-300
0-vibrations, 26, 36, 43, 168, 169, 230, high spin states, 280-282, 287, 290,
328-331 300-313
BILL, 352-356 off diagonal mixing, 280-312
binding energies, 7, 10, 29, 126, 127 rotation alignment, 267, 268, 294, 295,
300-313
Casimir operators, 202 unique parity states, 292-300
central potentials, 49-52 correlations (in ground state), 316, 320
371
372 Index

Coulomb excitation, 29, 335-344 geometric models, 34, 145-331

Coulomb force, 56, 57 g factors, 238, 239
Coulomb-nuclear interference, 343, 344 giant resonances, 146, 147, 232
cranked shell model, 312, 361 group theory (IBA), 199, 202-214
crossing frequency, 309, 310
cubic terms (in IBA), 220, 233 harmonic oscillator potential, 47, 50, 58
harmonic vibrator model, 42, 145-161
Davydov model, 168, 187-191 heavy ion fusion reactions, 29, 183, 351,
decoupled bands, 266, 279, 283, 284 356-361
deformation, 141-145, 156-331 heavy ion transfer reactions, 361-363
and configuration mixing, 141-145 and spin sensitivity, 363
intruder states, 156-161 hexadecapole deformation, 296-300, 344
m-shell distributions, 143-145
pairing, 101-103, 131 IBA, 143, 197-235
p-n interaction, 126, 141-145, and Alaga rules, 212-214, 228-230
156-161, 233-246 assumptions, 198, 199
p-p, n-n interactions, 126, 141-145, backbending, 233
156-161 bandmixing, 228-230
degenerate model (see RPA, TDA), collective models, 197, 198, 206-216,
320-322 220, 223, 226, 231
5-function residual interaction (see CQF, 217, 224-231
residual interactions), 69-104, dynamical symmetries, 199, 202-235
128-131 U(5), 205-208, 215, 217-220
detector arrays, 359 O(6), 205, 215-220
deuteron, 9, 14, 26 SU(3), 205, 208-215
dynamic deformation theory, 323 extensions to s-d IBA-1, 232, 233
dynamical symmetry, 199, 202-235 finite boson number, 206-209, 212,
218, 224, 227-231, 244, 245
energy, staggering in y-band, 194, 195
y-softness, 205, 215-221, 230-233
even-even nuclei, 69-108, 141-246
group structure, 199, 202-221
experimental techniques, 29, 59, 60, 64,
136, 137, 271-279, 288-299, NpNn, 231
symmetry triangle, 221-231
335-363
transition regions, 221-231
triaxiality, 220, 230, 231, 233
Federman-Pittel mechanism, 236-239
intruder states, 156-161, 239, 243-245
fermion symmetry models, 233
irreducible representations, 202
finite boson number effects, 206-209,
isospin, 12, 68, 69, 81-83, 141-145
212, 218, 224, 227-231, 244, 245
folds (coincidence), 359
/'-configurations, 67-108
f-spin (and multiplets), 232
/"-configurations, 104, 109-137
-vibrations, 36, 43, 166-169, 175-183, /^-configurations, 67-108
194, 195, 228-231, 323-326 jj coupling, 74
-ray multipolarities, 29
GAMS, 226, 227, 352-355 K quantum number, definition, 165, 166
gap (pairing), 80-85, 93, 132-137 K mixing (see Coriolis coupling),
gap parameter (A), 133 175-183, 191, 228-234
g-boson, 233
generator (group), 202, 203 ^-potential, 53, 263, 264
geometric interpretation of residual like-nucleon interaction, 10, 15, 34,
interactions, 76, 85-92, 96-103 59-62, 126, 127, 141-145, 156-161
 Index 373

Lip as formula, 185-186 Nilsson diagram, 250-260

LS-jj coupling, 74 Nilsson Hamiltonian, 256-269
orbit angle, 249-250
magic numbers, 8, 31, 48, 52 rotational motion, 264-270
Mikhailov plots, 178-181 single nucleon transfer, 271-279, 288,
and IBA, 228-230 290-299
mirror nuclei, 7, 10 small deformation limit, 257-260
valence mirror nuclei, 127 unique parity orbits, 252, 253
models U-,V-factors, 273, 279
algebraic, 3, 197-235 N Nn scheme (valence p-n interaction),
collective, geometric, 3, 34, 141-331 P
231,235-246
microscopic (RPA, TDA) 3, 24, 142, and collective models, 246
143,313-331 effective valence nucleon numbers,
Nilsson, 3, 33, 37, 247-312, 323-326 238, 239, 244
shell, 31, 33, 47-137 intruder states, 239, 243
moment of inertia, 131, 132, 164 monopole p-n interaction, 236-240
m-scheme, 109-111, 146, 147 nuclei far from stability, 240, 241
m-state distribution and deformation, odd A nuclei, 241
143-145 parametrization of collective models,
m-state localization, 72, 90, 96 241
multiparticle systems (shell model), P-factor, 242-246
109-137 Spn, 244, 245
multipole decomposition of residual saturation of collectivity, 244, 245
interactions, 59, 94-104, 236-240 subshell gaps, 236-240, 244
and limits on multipole order, 95-98 unified view of phase transitions, 245,
monopole interaction, 61, 98-100, 246
236-240 (n,n'7), 351
pairing, 94, 100, 104 nuclear force, 4, 6-17, 57
quadrupole interaction, 94,99-103 nuclear size, 3, 6
repulsion of like nucleons, 99 nuclei
repulsive effects of attractive interac- Ba, 34, 220, 303
tions, 99 Bi, 64, 66, 84, 92
multiphonon states, 148-156, 159-161, Ca, 31, 62-66, 125, 126
169, 170 Cd, 34, 35, 146, 156-160, 207, 239
multipolarities, 40 Cl, 84, 85, 95, 107
multistate mixing, 17, 22-25, 271-312, Dy, 36, 66, 191, 359
319-322, 327-330 Er, 36, 105, 174, 178, 179, 191, 266, 306,
and Coriolis coupling, 271-312 307, 325, 328, 349, 350
Gd, 174, 266, 342
(n, y), 344-356 Hf, 174, 191, 212-214, 261
n-hole configurations, 106, 107, 125 Hg, 243, 244
Nilsson model, 3, 33, 37, 145, 244, 245, K, 107
247-312, 319, 323-326 La, 266, 303
and asymptotic quantum numbers, Nb, 62, 65
253-256 Nd, 127-130, 363
Coriolis coupling, 265, 266, 279, Os, 191, 299, 300
280-312 Pb, 22, 32, 33, 64, 78, 105, 158, 239, 322
decoupled bands, 266, 279, 283, 284 Po, 33, 78
K splitting, 249, 250 Pt, 136, 137, 191, 218-220
large deformation limit, 258-260 Rb, 305, 306
374 Index

Re, 266 particle versus hole matrix elements,

Ru, 35, 161 106-108, 125
Sb, 64, 101 Pauli principle, 5, 14-17, 34, 46, 47, 56,
Sc, 44, 62, 101 57, 68, 69, 72-104, 109-111, 130,
Sm, 191, 363 144, 145
Sn, 33, 64, 66, 105, 146, 177, 239 and allowed J values, 68, 69
Sr, 239 independent particle motion, 47, 57
Te, 33-35, 78, 146, 160 m-scheme, 109-111
Ti, 114 pairing, 130
Th, 342 residual interactions, 72-104, 130
U, 266 valley of stability, 56
V, 114, 115 P-factor, 235-246
W, 36, 91, 262, 275-279, 288-299, 325 p-f bosons, 233
Xe, 34, 146, 220, 351 phase transitions, 196, 197, 221-231,
Yb, 33, 37, 66, 184, 212-214, 312 235-246
Zr, 62, 65, 66, 239 phonon model, 145-161, 169-170
nuclei far from stability, 240 potentials, 47
proton, neutron distinguishability, 81-83
p-n interaction, 10, 14-16, 34, 59-62,
oblate shape, 263, 264
81-85, 88-92, 95, 98-103, 127,
octupole vibrations, 36, 169, 328-330
141-145, 156-161, 233-246, 322
odd A nuclei, 247-312
and configuration mixing, 127,
odd particle blocking, 311, 312
141-145,322
orbit dependence of nuclear forces, 61,
deformation, 127, 141-145, 156-161
236-239,244,245
isospin formalism, 81-84
O(6), 193, 196, 205, 214-221
like-nucleon interactions, 10-15, 34,
59-62
pairing, 15, 76, 84, 93, 94, 100-104, mass dependence, 92
130-137,323 orbit dependence, 61, 236, 237, 244,
and deformation, 131, 132 245
^function interaction, 76, 93, 94, 100, pairing, 243
130, 131 parabolic rule, 99-103
ground state spin, 131 separation energies, 14, 15, 84
moment of inertia, 131, 132 strength of, 92, 244, 245
Nilsson orbits, 132 probe size, 95, 96
off-diagonal matrix elements, 93, 132, prolate, oblate shapes, 163, 263-264
135,136
pairing gap, 80, 93, 132-137 quadrupole deformation, 161-246
pairing property of residual interac- quadrupole interaction, 94-104
tions, 15, 104 quadrupole moments, 9, 14, 26, 107, 164,
parabolic rule, 101-103 171
quasi-particles, 132-136 quadrupole vibrations (spherical nuclei),
RPA,323 147-161
separation energies, 131 quasi-particles, 132-136, 323
transformation of energy zero, 134
pairing plus quadrupole (PPQ) model, range of nuclear force, 4, 10, 15-17,
94 96-103
parabolic rule, 99-103 and multipole order, 96-98
parity, 54, 55 repulsive aspect of like nucleon interac-
particle-hole spectra, 105, 107 tion, 29, 30, 80-84, 99, 126
 Index 375

residual interactions, 15,59,61,67-137 even tensors, 116-130

and binding energies, 126,127 odd tensors, 116-130
5-function interaction, 69-104 pairing property, 118
even tensors, 117, 121-124 separation of rotational and particle
geometric interpretation, 76, 85-92, motions, 247, 248
96-103 separation energies, 7, 13, 14, 29, 30, 131
multipole decomposition, 59, 94-104 shapes of nuclei, 29
odd tensors, 70, 95, 104, 116-120, shell model, 31, 33, 44-137
123-130 shell structure, 50-56
off-diagonal matrix elements, 108 single-particle energies, 50-54, 58-62,
pairing force, 15, 76, 93, 94, 100, 104, 98-100, 130
130-137 single-particle transfer reactions, 29, 59,
repulsive effects of attractive interac- 60, 64, 271-279, 288, 290-299, 326,
tions, 99 335
repulsion of like nucleons, 29, 30, 84, spectroscopic completeness, 348-351
99, 126 spherical-deformed transition regions,
seniority, 70, 104, 116-130 141-145, 196, 197, 221-231,
separation of radial and angular parts, 233-245
71, 83, 95 spin-orbit potential, 54-56, 263, 264
sensitivity to orbit, 61, 236-239 square well potential, 47
surface 5-interaction, 125, 129 subshell gaps, 236-240, 361
Ritz combinations, 351 surface 5-interaction, 125, 129
rotation alignment, 267, 268, 294, SU(3), 205, 207-214
300-313 superdeformation, 312, 359, 361
and Fermi surface, 300-310 symmetry triangle, 221-231
weak coupling, 303, 304 symplectic models, 233, 334
rotational energies, 183-186, 264-270 systematics, 28-43, 115-130, 164,
rotational motion, 36, 38, 165, 166, 176-181, 221-231, 233-246, 262,
264-270 326-331
rotation-vibration coupling, 175-183
RPA, 142, 146, 196, 313-331 T=0 p-n interaction, 81-93, 141-145, 236
and collectivity, 320-322, 327 T=l interaction, 81-93, 141-145, 236
degenerate model, 320-322 Talmi-Glaubman rule, 105
y-vibration, 323-326 TDA (see RPA), 143, 146, 313-331
HRPA, 323 transition rates and two-state mixing (see
Nilsson model, 319, 323-326 also B(E2) values, selection
schematic model, 318-322 rules), 25-27, 29
TDA, 313-331 transition regions, 141-145, 196, 197,
221-231,235-246
S'245 two-state mixing, 5, 17-27, 137, 181-183,
saturation of collectivity, 244, 245 187, 230, 250-253, 280-312,
scissors mode, 147, 232 327-330
s, d bosons (see IBA), 198 and Coriolis coupling, 280-312
second quantization, 148 level repulsion, 22
selection rules, 151, 160, 161, 170-183, multistate mixing, 17, 22-25, 327-330
189-193, 206, 207, 211, 212, strong mixing limit, 20, 21
217-221 weak mixing limit, 21, 22
seniority, 70, 104, 155-130
and conservation of, 117 U-factors, 133, 273-279
constancy of energies, 119 U(5), 205-207, 217-221
376 Index

valence particles, 33 VMI, 185

valley of stability, 29,30,56
V-factors, 133, 273-279 weak coupling and rotational alignment,
vibrational model, 42,145-161 303, 304
and B(E2) values, 42, 151-154, weakness of nuclear force, 14, 15, 57
159-161 Wilets-Jean model, 168, 187-193
energies 147-150, 154-161
multiphonon states, 147-161
vibrational motion, 26, 36, 43 yrast states, 357-361
vibrations of deformed nuclei, 166-168,
230-231,323-331 Z b .z b r ,Z r , 176-181