Theoretical Nuclear Physics

Chong Qi
[email protected]
Department of Physics, Royal Institute of Technology (KTH)

April 14, 2015

 Contents

1 Basic Quantum Mechanics concepts 11

1.1 The Hilbert space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.2 Eigenvalues and eigenvectors of matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Hermitian operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.3.1 Physical meaning of the eigenvectors and eigenvalues of Hermitian operators . . . 13
1.3.2 Representations and their use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4 Unitary Operator: symmetries and conservation laws . . . . . . . . . . . . . . . . . . . . 15
1.4.1 Translation symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4.2 Rotational symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4.3 Parity symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5 Sum of angular momenta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5.1 Symmetry properties of the Clebsch-Gordan coefficient . . . . . . . . . . . . . . . . 19
1.5.2 Short Algebraic Table of the 3-j Symbols . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.3 6-j symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.5.4 9-j symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.6 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2 Gamow states and the Berggren representation 23

2.1 One-particle Hamiltonian in one dimension . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2 Square barrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 Gamow states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.4 Berggren completeness relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3 Nuclear Shell Model 32

3.1 Introduction to the shell model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2 Shell model representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3 Shell model central potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4 Shell closures and the magic numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.4.1 Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.4.2 Particle excitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.4.3 Hole excitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.5 Parameterization of the Woods-Saxon Potential . . . . . . . . . . . . . . . . . . . . . . . . 43
3.5.1 Existing parameterizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.6 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4 Magnetic resonances in nuclei 49

4.1 Charge particles in a magnetic field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Time dependent magnetic fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2.1 Time-dependent perturbation treatment . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2.2 Rabi formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.3 Nuclear magnetic resonance (NMR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.3.1 Magnetic Resonance Imaging (MRI) . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.3.2 Electron spin resonance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.4 Magnetic fields and magnetic moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.4.1 Dipole magnetic moments in nuclei . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.4.2 Schmidt values of single-particle states . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.5 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

5 Rotational model 57
5.1 Rotational model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 Deformed shell model (Nilsson model) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.3 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
0 Contents 2

6 Two-particle states 63
6.0.1 Parity of n-particle states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.1 Isospin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.1.1 Ladder operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.1.2 Sum of isospins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.1.3 Consequences of antisymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.2 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

7 Second Quantization 69
7.1 Creation and annihilation operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7.1.1 Occupation number and anticommutation relations . . . . . . . . . . . . . . . . . . 69
7.2 Normal product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
7.2.1 Wick’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
7.3 One-body operator in second quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
7.4 Two-body operator in second quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7.5 Hartree-Fock potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7.6 Two-particle Random Phase Approximation (RPA) . . . . . . . . . . . . . . . . . . . . . . 75
7.7 Tamm-Dankoff Approximation (TDA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7.8 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

8 Shell model excitations 80

8.1 Two-particle energies and wave functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
8.2 Interaction matrix elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.2.1 The separable matrix element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
8.2.2 Pairing force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.3 Isospin symmetry in nuclei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.4 Nucleon-Nucleon Interaction and LS coupling . . . . . . . . . . . . . . . . . . . . . . . . . 85
8.4.1 General Aspects of the Nucleon-Nucleon interaction . . . . . . . . . . . . . . . . . 85
8.4.2 Yukawa theory and One-Pion Exchange potential . . . . . . . . . . . . . . . . . . . 86
8.4.3 LS coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
8.4.4 LS coupling in laboratory frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.4.5 Tensor force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.5 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

9 Nuclear pairing 92
9.1 Pairing gaps: odd-even binding energy differences . . . . . . . . . . . . . . . . . . . . . . . 92
9.1.1 Basic spectral properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
9.2 The seniority model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
9.3 Basics of pairing correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
9.3.1 Hartree-Fock-Bogoliubov theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
9.4 Exact solutions for pairing interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
9.4.1 The Richardson solution of the reduced BCS hamiltonian . . . . . . . . . . . . . . 99
9.4.2 Generalization to the Richardson-Gaudin class of integrable models . . . . . . . . 101
9.4.3 Applications of the Richardson solution to pairing in nuclear physics . . . . . . . . 101
9.4.4 The hyperbolic model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
9.5 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

10 Electromagnetic Transitions 107

10.1 The spin of the photon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
10.2 Electric and magnetic multipoles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
10.3 Applications to nuclear spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
10.4 Operators and transition rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.4.1 Electric transition operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.4.2 Magnetic transition operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.4.3 One-body transition operator in the coupled form . . . . . . . . . . . . . . . . . . 113
10.4.4 Moments in terms of electromagnetic operators . . . . . . . . . . . . . . . . . . . . 115
10.5 Nuclear matrix elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
10.5.1 Single-j configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
0 Contents 3

10.5.2 0+ → 0+ transitions. Internal and pair conversion . . . . . . . . . . . . . . . . . . 116

10.6 Weisskopf units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10.7 Deuteron Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10.7.1 Magnetic Moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
10.7.2 Quadrupole Moment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
10.8 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

11 Beta and alpha readioactive decays 123

11.1 Introduction to beta decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
11.1.1 Q-values in β − , β + and electron capture decays . . . . . . . . . . . . . . . . . . . . 124
11.1.2 Beta decay rates and the ft value . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
11.2 Fermi Theory of allowed beta decays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
11.2.1 logaritm of f t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
11.2.2 Fermi decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
11.2.3 Gamow-Teller decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
11.2.4 Effective operators for the Gamow-Teller decay . . . . . . . . . . . . . . . . . . . . 130
11.3 Sum rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
11.4 14 C-dating beta decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
11.5 Alpha decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
11.5.1 Thomas expression for the width . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
11.6 Homework problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

12 Nuclear energy and Nucleosynthesis 143

12.1 Fission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.1.1 Energy production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.1.2 Chain reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.1.3 Theory of nuclear fission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
12.2 Stability of superheavy elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.3 Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.3.1 Heavy-ion fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
12.4 Nucleosynthesis in stars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.4.1 Hydrogen burning: the pp chain and CNO cycle . . . . . . . . . . . . . . . . . . . 146
12.4.2 Helium burning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
12.4.3 The burning of heavy elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
12.4.4 The end of the star . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

13 Constants and units 152

 List of Figures

1 Experimental values of nuclear binding energies per nucleon (in MeV) for all known nuclei
from Ref. [2]. The three cases with binding energies per nucleon less than 2 MeV cor-
respond to 4 H (1.720 MeV), 4 Li (1.154 MeV) and 2 H (1.112MeV). In comparison, 4 He is
bound by as much as 7.074 MeV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Stability of nuclides as a function of N and Z from Ref. [1]. . . . . . . . . . . . . . . . . . 9

1.1 Angle δϕ corresponding to the rotation of the system and the relation among the radius
~r, the radius increment ~a and the angle increment δϕ. . . . . . . . . . . . . . . . . . . . . 17

2.1 Square well potential in one-dimension. The range of the potential is a and the depth is
−V0 . For x < 0 the potential is infinite and, therefore, the wave function vanishes at x = 0.
Eb (Ec ) is the energy of a bound (continuum) state. . . . . . . . . . . . . . . . . . . . . . 24
2.2 Radial function of a narrow resonance and a bound state. The solid and the dashed line
denote the real and imaginary part of the wave function of a narrow resonance respectively,
while the dotted line denotes the wave function of a bound state. . . . . . . . . . . . . . . 27
2.3 Radial function φ(r) corresponding to the single-particle neutron antibound state 0s1/2 at
an energy of -0.050 MeV. Taken from Ref. [1]. . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4 A schematic picture of the halo nucleus 11 Li. . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.5 Integration contour L+ in the complex energy plane. The open circles denote the reso-
nances included in the sum of Eq. (2.14), while the solid circles are those excluded. The
vertex (c, 0) corresponds to the energy cutoff point c. . . . . . . . . . . . . . . . . . . . . . 29

3.1 Potential describing bound single-particle states. The nuclear central field is well approxi-
mated by the Woods-Saxon potential (3.18). In the region where the single-particle states
lie the nuclear (Woods-Saxon) potential can be approximated by an Harmonic oscilla-
tor potential. Close to the continuum threshold the nuclear potential vanishes and the
centrifugal plus Coulomb potentials become dominant. . . . . . . . . . . . . . . . . . . . 36
3.2 In the left is the energy spectrum corresponding to a Harmonic oscillator potential of
frequency ω. In the middle is the same plus a term Dl2 , with D negative. The spectrum
to the rigth corresponds to the shell model Hamiltonian (3.5), which follows very well the
tendencies of the experimental data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3 Single particle states in the shell model potential. The energies  are measured from the
continuum threshold. The Fermi level is indicated as FL. . . . . . . . . . . . . . . . . . . 41
3.4 Experimental spectrum of the nucleus 57 Ni29 taken from the Internet site mentioned in the
text. Energies are in MeV. The level at 3.010 MeV has not been completely determined
yet. It can be 7/2+ as well as 9/2+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.5 Excitations hi in the (A-1)-particle nucleus. It looks like at the level h1 there is a hole in
the completely filled states of the core. Therefore the states below the Fermi level are called
”hole excitations”. A hole at the more deeply bound level h2 induces a higher excitation
than the one at h1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.6 Experimental spectrum of the nucleus 131 50 Sn81 . The levels are in parenthesis, meaning that
they are not completely determined yet. The level 11/2− is at 0.0+x MeV, implying that
it is only slightly above the 3/2+ ground state. . . . . . . . . . . . . . . . . . . . . . . . . 44

4.1 The resonant form of the signal as the energy E, corresponding to the weak magnetic field
B1 , approaches the energy E0 induced by B0 . The width of the resonance is Γ. . . . . . 52

5.1 A cylindrically symmetric rotator. The symmetry axis is z 0 . The projection of the angular
momentum upon this axis is K, and upon the z-axis in the laboratory frame is M . . . . 58
5.2 Ground rotational band (i. e. K=0) of 238 P u. Energies are in keV. The rotational energies
E(J, 0) = E(2, 0) J(J + 1)/6, where E(2, 0) = 44keV in this case, are E(4, 0)=147 keV,
E(6, 0)=308 keV and E(8, 0)=528 keV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.3 Distribution of nuclei with respect to deformation indicator R42 . Taken from GF Bertsch,
arXiv:1203.5529. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
0 List of Figures 5

5.4 The different nuclear shapes that can be parametrised by spherical harmonic functions,
where λ characterises the different orders of the corresponding distributions. . . . . . . . 60

8.1 Nuclear spectra of the nuclei 18 18 18

10 O8 , 9 F9 and 8 O10 . . . . . . . . . . . . . . . . . . . . . . 84
8.2 The spectra of mirror nuclei Ni, Co and 54 Fe (with T = 1).
54 54
From PRL 97, 152501
(2006). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
8.3 The spectra of mirror nuclei 48 Mn and 48 V (with T = 3/2). From PRL 97, 132501 (2006). 85

9.1 Upper panels: odd-N pairing gaps. Lower panels: even-N pairing gaps. Typically, the
odd-N nuclei are less bound than the average of their even-N neighbors by about 1 MeV.
However, one sees that there can be about a factor of two scatter around the average value
at a given N . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
9.2 Upper panels: odd-Z pairing gaps. Lower panels: even-Z pairing gaps. . . . . . . . . . . . 93
9.3 Energy levels of odd-N Sn isotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
9.4 Energy gap in the excitation spectrum of even-even nuclei, scaled to 2∆(3) . See text for
details. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

10.1 Relation between the systems of coordinates related by a rotation θ along the axis z = z 0 . 108
10.2 The first four levels in the nucleus 150 Er with the measured gamma ray transitions. The
task is to determine the spins and parities of the levels a, b and c. . . . . . . . . . . . . . 111
10.3 Radial wave functions R0 (r) and R2 (r) calculated from realistic N N interactions. . . . . . 118
10.4 Schematic picture of the relative orientation of ẑ and r̂. . . . . . . . . . . . . . . . . . . . 120

11.1 Unstable state at energy E from where the alpha particle is emitted. The penetrability is
determined by the width of the barrier at the energy of the state, indicated by a dashed
line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
11.2 Beta decays of the odd-odd nuclei (N, Z) and (N − 2, Z + 2). . . . . . . . . . . . . . . . 140
11.3 Beta decay of the odd-odd nucleus (N, Z). . . . . . . . . . . . . . . . . . . . . . . . . . . 140
11.4 Alpha decay scheme of the even-even nucleus (N + 2, Z + 2). . . . . . . . . . . . . . . . . 140
 List of Tables

3.1 Energy levels corresponding to an Harmonic oscillator potential of frequency ω. The states
are labelled by N = 2n + l. The energies Enl are in units of ~ω. Dl is the degeneracy of
the state (n, l). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.2 Commonly used Woods-Saxon parameter sets in the literature. . . . . . . . . . . . . . . . 46

4.1 the gyromagnetic ratio γ of typical nuclei. . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

8.1 The four LST combinations that are compatible with the Pauli principle. . . . . . . . . . 88

11.1 Values of log10 f t and the corresponding restrictions upon the angular momentum transfer
∆J and parity change ∆π. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
||s||jb i
11.2 Gamow-Teller reduced single-particle matrix elements 2hja√ 3
. . . . . . . . . . . . . . . 130
JT
11.3 Comparison between diagonal matrix elements hij|V |iji (in MeV) of empirical and re-
alistic interactions. The realistic interactions are calculated from the CD-Bonn and N3 LO
potentials with the G-matrix (G) and Vlow−k (K) renormalization approaches. Only the
few terms that related to the description of the dating β decay are listed for simplicity. . . 133
11.4 Comparison between different wave functions calculated with empirical and realistic inter-
actions. All calculations are done with the code [26] except those of Jancovici and Talmi’s
and of the chiral potential which are taken from Ref. [7] and Ref. [13], respectively. . . . . 134
11.5 Same as Table 11.3 but for the non-diagonal matrix elements hij|V |kliJT (in MeV) of
empirical and realistic interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
11.6 The central (C), spin-orbit (SO) and tensor (T) components of the matrix elements hij|V |kliJT
of empirical and realistic interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
11.7 Wave functions of 14 C and 14 N calculated with the central force and central and spin-orbit
force components of empirical and realistic effective interactions. . . . . . . . . . . . . . . 137

13.1 The values in SI units of some non-SI units based on the 2010 CODATA adjustment of
the values of the constants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
13.2 The values of some energy equivalents derived from the relations E = mc2 = hc/λ = hν =
kT ; 1 eV = (e/C) J, 1 u = mu = 12 1
m(12 C) = 10−3 kg mol−1/NA . . . . . . . . . . . . . . . 153
13.3 The values of some energy equivalents derived from the relations E = mc2 = hc/λ = hν =
kT ; 1 eV = (e/C) J, 1 u = mu = 12 1
m(12 C) = 10−3 kg mol−1/NA . . . . . . . . . . . . . . . 154
13.4 Some exact and measured quantities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
13.4 (Continued). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
 Introduction

The atomic nucleus is a quantum man-body object composed of A nucleons (mass number): Z protons
(atomic number) and N neutrons. An atomic species with the specified numbers A and Z is also often
called a nuclide [1]. One denotes by AZ XN the nucleus X with the total number of nucleons A = N + Z
(although the simpler notation A X is often used). Such a system is stable only for certain combinations
of numbers Z and N . Presently, around 300 stable nuclides are known. Systems different from stable
configurations undergo spontaneous radioactive decays until the stability is reached. A nucleus of such an
unstable nuclide is considered as a well defined object if its half-life is much longer than 10−21 s which is
a characteristic timescale for processes governed by strong interaction. These nuclides are bound by the
nuclear force and/or by Coulomb and centrifugal barriers. The number of unstable nuclides synthesized
in laboratories is constantly growing, and up to now more than 3000 were identified.
Nomenclature
• Nuclide: A specific nuclear species, with a given proton number Z and neutron number N
• Nucleon: Neutron or proton
• Isotopes: Nuclides of same Z and different N
• Isotones: Nuclides of same N and different Z
• Isobars: Nuclides of same mass number A (A = Z + N)
• Isomer: Nuclide in an excited state with a measurable half-life
• Mesons: Unstable particles composed of one quark and one antiquark. Some light mesons are the
associated quantum-field particles that transmit the nuclear force.
A brief history of nuclear (structure) physics:
• 1896: Discovery of radioactivity (Becquerel)
• 1911: Discovery of the nucleus (Rutherford)
• 1932: Discovery of the neutron (Chadwick)
• 1935: Bethe-Weiszaker mass formula
• 1939: Discovery of (neutron-induced) fission
• 1949: Shell model (Goeppert-Mayer, Jensen)
• 1951: Collective model (Bohr, Mottelson, Rainwater)
• 1957: Nuclear superfluidity (Bohr, Mottelson)
• Since then: Nuclear forces, many-body methods (HF, HFB, RPA, GCM, Green function, etc).
It is very difficult to treat the nuclear many-body system. On the one hand, the nucleon-nucleon
interaction provided by fundamental theories, as Quantum Chromo Dynamics (QCD), is extremely dif-
ficult to apply. On the other hand the many-body problem itself is very difficult since the number of
particles is large, but not large enough to be able to be treated in statistical terms, as it happens with
other many-body systems like e. g. condensed matter. Therefore the main tasks in nuclear structure
studies is first to find effective forces that explain the available experimental data and then to be able
to perform this task within a manageable theoretical framework. In this course we will present the most
important solutions that have been found to perform those tasks.
Effective forces are introduced to explain in a treatable fashion nuclear properties. Thus, it is known
that all nuclei with N even and Z even (even-even nuclei) have in the ground state spin and parity 0+ .
This indicates that nucleons of the same kind are arranged in pairs. The nucleons in each pair are coupled
with the corresponding spins pointing in opposite directions, such that the pair is coupled to zero angular
momentum. The force inducing this pairing of nucleons is called ”pairing force”.
0 List of Tables 8

The nucleus is an isolated system with a well defined total angular momentum. It is common practice
to represent this total angular momentum of a nucleus by the symbol I (or J) and to call it nuclear spin.
Associated with each nuclear spin is a nuclear magnetic moment which produces magnetic interactions
with its environment. It should be noted that for electrons in atoms we make a clear distinction between
electron spin and electron orbital angular momentum and then combine them to give the total angular
momentum.

Figure 1: Experimental values of nuclear binding energies per nucleon (in MeV) for all known nuclei from
Ref. [2]. The three cases with binding energies per nucleon less than 2 MeV correspond to 4 H (1.720
MeV), 4 Li (1.154 MeV) and 2 H (1.112MeV). In comparison, 4 He is bound by as much as 7.074 MeV.

The atomic mass Mat , the nuclear mass Mnu and the electron mass M el are related by

Mat = Mnu + ZMel − Bel /c2 (1)

where Z is the number of electrons (and therefore of protons in a non-ionized atom) and Bel is the
binding energies of the electrons (notice that Bel is subtracted). It is therefore straightforward, by using
the above relation, to obtain the nuclear mass from the atomic mass. Usually what is given not the
measured atomic masses but what is called the ”mass excess”, which is defined by

∆M = Mat − A · amu (2)

where A = N + Z is the atomic number and Mu is the atomic mass unit, which is defined as one-twelfth
of the mass of a neutral 12 C atom.
The convenient unit for measuring the nuclear mass is called the atomic mass unit or for short amu.
The mass of a 12 C atom (including all six electrons) is defined as 12 amu. We have

1amu = 1.6605402(10) × 10−27 kg = 931.49432(28)M eV /c2 . (3)

The total binding energy B(N, Z) is defined as the total minimum work that an external agent must
do to disintegrate the whole nucleus completely. By doing so the nucleus would no longer be existent but
disintegrated into separated nucleons. This can also be considered as the total amount of energy released
when nucleons, with zero kinetic energy initially, come close enough together to form a stable nucleus.
The B(N, Z) is the binding energy of the nuclide related to its mass M (N, Z) as

M (N, Z) = Z MH + N mn − B(N, Z), (4)

where MH and mn are masses of the hydrogen atom and the neutron, respectively.
Another important nuclear quantity is the binding energy per nucleon, which for some nucleon num-
bers N and Z is larger than the average. These numbers are called ”magic numbers”. They include
0 List of Tables 9

stable
1014 yr

160 1012 yr

1010 yr
140
108 yr

106 yr
120
104 yr

100 yr
100
1 yr
Z=N 106 s
80
104 s

60 100 s

1s

40 10−2 s

10−4 s
20
10−6 s

10−8 s
N unstable
Z 20 40 60 80 100

Figure 2: Stability of nuclides as a function of N and Z from Ref. [1].

the numbers 2, 8, 20, 28, 50, 82 and 126. A nucleus having magic N and Z, like e. g. 16 8 O8 (Oxygen
16), 48
20 Ca28 (Calcium 48) and 208
82 P b126 (Lead 208) are very bound. It was found that a nucleon moving
outside the nuclear field induced by a double magic nucleus does not influence appreciably the motion of
the nucleons inside the nucleus. The double magic nucleus is as a ”frozen core”. The field induced by the
core acts as a whole upon the odd nucleon moving outside the even-even core. In other words, the low-
lying excitations of even-odd or odd-even nuclei outside a magic core can be considered as single-particle
excitations.
The proton- and the neutron separation energy of a nuclide with numbers N and Z are given by
Sp (N, Z) = B(N, Z) − B(N, Z − 1) (5)
Sn (N, Z) = B(N, Z) − B(N − 1, Z). (6)
When we move along the line of isotopes with the given atomic number Z, starting from stability
towards neutron-deficient nuclides, the proton separation energy Sp decreases and at certain location it
becomes negative. The proton drip-line is defined as the border between the last proton-bound isotope
and the first one with the negative value of the Sp . In the fully analogous way, the neutron drip-line for
a given neutron number N is defined as a border between the last neutron bound isotone, when counting
from stability, and the first one for which the neutron separation energy Sn is negative. The drip lines as
defined above are very useful in identifying and discussing limits of stability. The two-nucleon separation
energies are given as
S2p (N, Z) = B(N, Z) − B(N, Z − 2) (7)
S2n (N, Z) = B(N, Z) − B(N − 2, Z). (8)
The separation energy of an α particle is given as
Sα (N, Z) = B(N, Z) − B(N − 2, Z − 2) − B(2, 2). (9)
Most heavy nuclei with masses A > 150 are unstable against α emission with negative Sα (or positive
Qα ). The Qα for α decay is given as
Qα (N, Z) = B(N − 2, Z − 2) + B(2, 2) − B(N, Z). (10)
 Bibliography

[1] http://en.wikipedia.org/wiki/Nuclide
[2] G. Audi, M. Wang, A.H.Wapstra, F.G.Kondev, M.MacCormick, X.Xu, and B. Pfeif-
fer, Chinese Phys. C 36, 1157 (2012); M. Wang, G. Audi, A. Wapstra, F. Kon-
dev, M. MacCormick, X. Xu, and B. Pfeiffer, Chinese Phys. C 36, 1603 (2012).
http://amdc.impcas.ac.cn/evaluation/data2012/data/mass.mas12
Chapter 1
Basic Quantum Mechanics concepts

Hermitian operators. Physical meaning of the eigenvectors and eigenvalues of Hermitian operators. Rep-
resentations and their use. Unitary Operator: symmetries and conservation laws. Sum of angular mo-
menta. 3j, 6j and 9j symbols.

1.1 The Hilbert space

In Quantum Mechanics the states are represented by vectors in an abstract space called Hilbert space.
Thus, a state α is a vector which, in Dirac notation, is written as |αi. As we will see below, this vector can
be associated either to a function Ψα (~r), which is regular and square integrable, or to a one-dimensional
matrix (spinor). In the first case the metric of the space is defined by the scalar product in the region V
of the three-dimensional (physical) space where the functions are square integrable. Usually V includes
the whole space. Thus, the scalar product between the vectors |αi and |βi is defined by
Z
hβ|αi = d~r Ψ∗β (~r)Ψα (~r) (1.1)
V

For the case of an N-dimensional spinor the vector α is associated to the one-dimensional matrix given
by  
α1

 α2 

 . 
  (1.2)

 . 

 . 
αN
and the scalar product between the vectors α and β is given by
 
α1
 α2 
  XN
.
β1∗ , β2∗ , ∗
βi∗ αi

 
hβ|αi = ., ., ., βN  = (1.3)

 . 
 i=1
 . 
αN

The vector hα| is called ”bra” and |αi is called ”ket”. The scalar product hα|βi is called ”bracket”.
∗
p of the scalar product that it is hα|βi = hβ|αi . Therefore
One sees from the definition the norm Nα
of a vector |αi, i. e. Nα = hα|αi is a real number. In Quantum Mechanics Nα2 is the probability of
measuring the system in the state α. Since the system exists, this probability should be Nα2 = 1. Notice
that we assume that the system is stationary, that is all processes are time-independent. Therefore if the
system is in the state α, it will remain there for ever. Below we will describe this system in terms of a
set of vectors |ni. The probability of measuring the system in the state |ni is |hα|ni|2 .

1.2 Eigenvalues and eigenvectors of matrices

The word ”eigenvector” almost always means a ”’right column” vector that must be placed to the ”right”
of the matrix A. The eigenvalue equation for a matrix A is

Av − λv = 0 (1.4)
1 Basic Quantum Mechanics concepts 12

which is equivalent to
(A − λI)v = 0. (1.5)
There may be also single-”row” vectors that are unchanged when they occur on the ”left” side of a
product with a square matrix A,
uA = λu. (1.6)
The left eigenvectors of A, are transposes of the right eigenvectors of the transposed matrix AT ,

AT uT = λuT . (1.7)

In matrix form we can write,

    
A1,1 A1,2 ... A1,n v1 v1
 A2,1 A2,2 ... A2,n 
  v2 
   v2 
..   ..  = λ  ..  (1.8)
  
 .. .. ..
 . . . .  .  .
An,1 An,2 ... An,n vn vn

where for each element we have,

n
X
wi = Ai1 v1 + Ai2 v2 + · · · + Ain vn = Aij vj = λvi . (1.9)
j=1

The inverse of a matrix is given as

AA−1 = A−1 A = I, (1.10)

where I is the identity matrix. For a simple 2 × 2 matrix we have

 
1 A2,2 −A1,2
A−1 = . (1.11)
A11 A2,2 − A1,2 A2,1 −A2,1 A1,1

The determinant of a 2 × 2 matrix A is given as

det(A) = A11 A2,2 − A1,2 A2,1 . (1.12)

The trace pf a n × n matix is given as

n
X
tr(A) = Aii . (1.13)
i=1

For a 2 × 2 matrix the eigenvalues are found to be

p
tr(A) ± tr2 (A) − 4det(A)
λ= . (1.14)
2
For the eigen vectors we have
v1 A12 λ − A22
= = . (1.15)
v2 λ − A11 A21

1.3 Hermitian operators

An operator A acting upon a vector |αi in the Hilbert space converts this vector into another one |βi.
It is important to point out that the Hilbert space we consider is closed, that is all vectors belong to the
space. In the applications that we will encounter in the course of these lectures only small subspaces of
the total Hilbert space (which usually has infinite dimension) will be chosen. In such a case the operator
 may bring |αi to a vector |βi lying outside the subspace. But we will not treat such situations here.
In other words, the systems we will treat are always closed.
The operator  is called Hermitian if

† =  Hermitian (1.16)

1 Basic Quantum Mechanics concepts 13

The eigenvalues a and eigenvectors |αi of the operator  satisfy the equation

Â|αi = a|αi (1.17)

and for the adjoint operator A† it is,

hα|† = a∗ hα| (1.18)
If the operator  is Hermitian one has

hβ|Â|αi = hβ|† |αi (1.19)

If, in addition, |αi and |βi are eigenvectors of Â, then one gets

ahβ|αi = b∗ hβ|αi (1.20)

Which implies that 

|αi =
6 |βi; hβ|αi = 0
(1.21)
|αi = |βi; a real
That is, the eigenvectors of a Hermitian operator form an orthogonal set and the corresponding eigenvalues
are real. These two properties play a fundamental role in Quantum Mechanics, as we repeatedly will see in
the course of these lectures. In quantum mechanics, the expectati on value of any physical quantity must
be be real and therefore an operator that corresponds to a physical observable must also be Hermitian.

Physical meaning of the eigenvectors and eigenvalues of Hermitian operators

The most important property of Hermitian operators in Quantum Mechanics is that their eigenvalues are
real. This property has allowed one to interpret these operators as the devices used to measure physical
quantities. One postulates that the Hermitian operator represents the apparatus used to measure a
physical quantity and the corresponding eigenvalues are all the possible values that one can obtain from
the measurement. In other words, only those values are allowed and nothing else. This is a radical
departure from Classical Mechanics, where one can give any value one wishes to all physical quantities
(for instance the energy).
The eigenvectors of Hermitian operators are the corresponding wave functions that allow one to
evaluate all probabilities, in particular transition probabilities. Besides, they play a fundamental role in
Quantum Mechanics. Thus, normalizing the eigenvectors in Eq. (1.20) as hα|αi = 1 and from Eq. (1.21)
one finds that they satisfy
hα|βi = δαβ (1.22)
which means that they form an orthonormal set of vectors in the Hilbert space. They can be used as
a basis to describe any vector belonging to the space. In a more rigorous statement one can say that
the eigenvectors of an Hermitian operator span the Hilbert space on which the operator acts. To see the
great importance of this property, assume a Hilbert space of dimension N and a Hermitian operator Â
acting on this space such that,

Â|αi i = ai |αi i, i = 1, 2, · · · , N (1.23)


Any vector |vi in the space spanned by the basis |αi can be written as

N
X
|vi = ci |αi i (1.24)
i=1

From Eq. (1.22) one obtains

ci = hαi |vi (1.25)
The numbers hαi |vi are called ”amplitudes”. If the vector |vi represents a physical (quantum) state, then
the amplitudes have to obey the normality relation given by,
N
X N
X N
X
∗
hv|vi = ci hv|αi i = hv|αi i hv|αi i = |hv|αi i|2 = 1 (1.26)
i=1 i=1 i=1
1 Basic Quantum Mechanics concepts 14

From Eq. (1.25) the vector |vi can be written as,

N
X N
X
|vi = hαi |vi|αi i = |αi ihαi |vi (1.27)
i=1 i=1

which shows that

N
X
|αi ihαi | = Iˆ (1.28)
i=1

This is the projector into the space spanned by the set |αi . We will use the projector often in these
lectures.

Representations and their use

One of the most important problems in theoretical nuclear physics is to evaluate the eigenvectors and
eigenvalues of a given operator B̂. That is, to find the vectors β and numbers b defined by

B̂|βi = b|βi (1.29)

This is called the ”Eigenvalue problem”. To evaluate the eigenvectors and eigenvalues one first chooses
a basis, that is a set of orthonormal vectors |αi which are usually provided by the diagonalization of a
Hermitian operator. This basis is also called ”Representation”,
 for reason which will become clear below.
If the number of vectors forming the orthonormal set |αi , i. e. the dimension of the basis, is N one
has, applying Eq. (1.28)
N
X N 
X 
B̂|βi = b|βi = B̂ |αi ihαi |βi =⇒ hαj |B̂|αi i − bδij hαi |βi = 0 (1.30)
i=1 i=1

which provides the equation to evaluate the amplitudes as,

N 
X 
hαj |B̂|αi i − bδij hαi |βi = 0 (1.31)
i=1

This is a set of N × N homogeneous linear equations in the N unknowns amplitudes hαi |βi. Besides the
trivial solution hαi |βi=0 for all i, one finds the physical solutions by requiring that the equations (1.31)
are linearly dependent upon each other. This occurs if the corresponding determinant vanishes. That is

hαj |B̂|αi i − bδij =0 (1.32)

which allows one to calculate N values of b. To calculate the amplitudes one disregard one of the Eqs.
(1.31) and the remaining N − 1 equations plus the normalization condition given by
N
X
|hαi |βi|2 = 1 (1.33)
i=1

give a non-linear
 N × N set of equations from which the amplitudes hαi |βi are extracted.
The set |αi can be a continuum set. An example of this is the eigenvectors corresponding to the
distance operator, i. e.
r̂|ri = r|ri (1.34)
The operator r̂ represents the device used to measure the distance (a rule for instance), |ri is the
corresponding vector in the Hilbert space and r the length one measures. Since

r̂ = r̂† (1.35)

one gets the projector as Z

dr|rihr| = Iˆ (1.36)
1 Basic Quantum Mechanics concepts 15

where, in spherical coordinates, it is r = (r, θ, ϕ) and dr = r2 dr sin θdθdϕ. One cannot speak of a number
of dimensions of this continuous basis, since it comprises all real numbers (which cannot be labeled by
integers). Therefore one uses the name ”representation” for the projector (1.36). In r-representation the
eigenvalue problem is Z
B̂|βi = b|βi = B̂ dr 0 |r 0 ihr 0 |βi (1.37)
Z
B̂ dr 0 hr|r 0 ihr 0 |βi = bhr|βi (1.38)

With hr|r 0 i = δ(r − r 0 ) and Ψβ (r) = hr|βi, one gets

B̂Ψβ (r) = bΨβ (r) (1.39)

and the matrix elements can readily be evaluated as,

Z Z Z
hαj |B̂|αi i = dr dr hαj |rihr|B̂|r ihr |αi i = drΨ∗j (r)B̂Ψi (r)
0 0 0
(1.40)

1.4 Unitary Operator: symmetries and conservation laws

An operator is Unitary if its inverse equal to its adjoints

U −1 = U † , (1.41)

or
U U † = U † U = 1. (1.42)
Unitary operators are of a fundamental importance to describe transformations of the system (basis).
We will analyze in this Lecture the cases of translations, rotations and parity.

Translation symmetry

The translation operator T̂ is defined by,

T̂ (∆r)|ri = |r + ∆ri (1.43)

which applied to the vector |Ψi gives,

Z Z
T̂ (∆r)|Ψi = dr T̂ (∆r)|r ihr |Ψi = dr 0 |r 0 + ∆rihr 0 |Ψi
0 0 0
(1.44)

and, in r-representation, the translated function is

Z
Ψt (r) = hr|T̂ (∆r)|Ψi = dr 0 δ(r − r 0 − ∆r)hr 0 |Ψi = Ψ(r − ∆r) (1.45)

Since
hr|T̂ † (∆r)T̂ (∆r)|ri = hr + ∆r|r + ∆ri = 1 (1.46)
one obtains
T̂ † T̂ = 1 (1.47)
which defines the operator T̂ as unitary.
The invariance of a wave function with respect to translations implies the conservation of the linear
momentum, This can be seen by noticing that the time dependence of an operator  is given by

d ∂  i
= + [H, Â] (1.48)
dt ∂t ~
Assume a system for which
HΨn (x) = En Ψn (x) (1.49)
1 Basic Quantum Mechanics concepts 16

If there is translation invariance, then

HΨn (x + ∆x) = En Ψn (x + ∆x) (1.50)

Making a Taylor expansion of Ψn one gets,

∞ ∞ k
(∆x)k dk Ψn (x) X 1

X d d
Ψn (x + ∆x) = = ∆x Ψn (x) = e∆x dx Ψn (x) (1.51)
k! dxk k! dx
k=1 k=1

and defining the linear momentum operator in the usual fashion as,
~ d
px = (1.52)
i dx
one obtains,
i
Ψn (x + ∆x) = e ~ ∆xpx Ψn (x) (1.53)
Therefore the translation operator is
i
T̂ (∆x) = e ~ ∆xpx (1.54)
and one has
H T̂ (∆x)Ψn (x) = HΨn (x + ∆x)
= En Ψn (x + ∆x) (1.55)
= En T̂ (∆x)Ψn (x) = T̂ (∆x)HΨn (x)

which implies,
[H, T̂ ] = 0 (1.56)
i. e.
[H, px ] = 0 (1.57)
Since px is time independent it is ∂ pˆx /∂dt =0 and from Eq. (1.48) one gets dpˆx /dt=0, which means that
the linear momentum is conserved.

Rotational symmetry

Performing a rotation of the system by an angle δϕ, as shown in Fig. 1.1, a function Ψ(r) is transformed
to Ψ(r + a). As seen in the Figure, it is a = δϕ rsinθ and the relation among the vectors a, δϕ and r is

a = δϕ × r (1.58)

Calling
F (r) = Ψ(r + a) (1.59)
one gets
Ψ(r) = F (r − a) = F (r) − a · ∇F (r) + · · · (1.60)

Performing a Taylor expansion as,

F (x + ∆x, y + ∆y, z + ∆z) =

∂F (x, y, z) ∂F (x, y, z) ∂F (x, y, z)
F (x, y, z) + ∆x + ∆y + ∆z + · · · (1.61)
∂x ∂y ∂z
one gets,
F (r + ∆r) = F (r) + ∆r · ∇F (r) + · · · (1.62)
With δϕ infinitesimal it is,

Ψ(r) − F (r) = −(δϕ × r) · ∇F (r) = −δϕ · (r × ∇F (r)) (1.63)

and replacing momenta by the corresponding operators one gets,

~ i i
p= ∇ =⇒ r × ∇ = r × p = L (1.64)
i ~ ~
1 Basic Quantum Mechanics concepts 17

rsi
nθ δφ
a
r

y
φ
δφ
x

Figure 1.1: Angle δϕ corresponding to the rotation of the system and the relation among the radius ~r,
the radius increment ~a and the angle increment δϕ.

Finally one obtains,  

i i
Ψ(r) = F (r) − δϕ · LF (r) = 1 − δϕ · L F (r) (1.65)
~ ~
For a finite angle ϕ one defines a small angle by using δϕ = ϕ/n, where n is a large number. Rotating
n times, i. e. applying the rotation operator n times, and in the limit of n = ∞, one gets,
 n
i ϕ·L
Ψ(r) = lim 1 − F (r) (1.66)
n→∞ ~ n

Since the Euler’s number e is

n
e = lim (1 + 1/n) (1.67)
n→∞

one can write

i i
Ψ(r) = e− ~ ϕ·L F (r) = e− ~ ϕ·L Ψ(r + a) (1.68)
Therefore the rotation operator is
i
UR = e− ~ ϕ·L (1.69)
and one gets
Ψ(r + a) = UR−1 Ψ(r) (1.70)
If the Hamiltonian is rotational invariant

HΨn (r + a) = En Ψ(r + a) = En UR−1 Ψ(r) = UR−1 HΨ(r) (1.71)

UR HΨn (r + a) = UR HUR−1 Ψ(r) = HΨ(r) (1.72)

[H, UR ] = 0 =⇒ [H, L] = 0 (1.73)
That is, if there is rotational invariance, then the angular momentum is conserved.
1 Basic Quantum Mechanics concepts 18

Parity symmetry

The parity operator π̂ is defined as,

π̂|xi = | − xi (1.74)
†
it is π̂ = π̂ (exercise) The eigenvalues of the parity operator are obtained as,

π̂|Ψλ i = λ|Ψλ i =⇒ π̂ 2 |Ψλ i = λ2 |Ψλ i (1.75)

since
π̂ 2 |xi = |xi (1.76)
one gets
λ2 |Ψλ i = |Ψλ i =⇒ λ = ±1 (1.77)
in x-space it is
hx|π̂|Ψλ i = h−x|Ψλ i = Ψλ (−x) = λΨλ (x) (1.78)

even, λ = 1
Ψλ (x) = (1.79)
odd, λ = −1
If [H, π̂]=0, as it happens with potentials with reflection symmetry, parity is conserved and λ is a good
quantum number.

1.5 Sum of angular momenta

We will first analyze the possible angular momentum values of a two-particle system. The angular
momenta of the particles are l1 and l2 and the total angular momentum is L = l1 + l2 . The components
L̂x , L̂y , L̂z of L satisfy the commutation relations

[L̂2 , L̂i ] = 0 (i = x, y, z) (1.80)

[L̂x , L̂y ] = i~L̂z , [L̂y , L̂z ] = i~L̂x , [L̂z , L̂x ] = i~L̂y (1.81)
and the same for l1 and l2 .
Besides, since the degrees of freedom of the particles are independent of each other one also has
2 2
[l̂1 , l̂2 ] = 0, [lˆ2 , l̂1 ] = [lˆ1 , l̂2 ] = 0. (1.82)

The eigenvectors corresponding to these operators are given by

L̂21 |l1 m1 i = ~2 l1 (l1 + 1)|l1 m1 i ; L̂1z |l1 m1 i = ~m1 |l1 m1 i

L̂22 |l2 m2 i = ~2 l2 (l2 + 1)|l2 m2 i ; L̂2z |l2 m2 i = ~m2 |l2 m2 i (1.83)
L̂2 |lmi = ~2 l(l + 1)|lmi ; L̂z |lmi = ~m|lmi

|l1 − l2 | 6 l 6 l1 + l2 , m = m1 + m2 . (1.84)
We have
−li 6 mi 6 li , −l 6 m 6 l (1.85)
Not all the quantum numbers related to these operators can be used to label the states. In other
words, not all of them can be taken as good quantum numbers. To see the reason for this we will analyze
the behavior of commuting operators.
Given two operators and their eigenstates as

Â|αi = a|αi and B̂|βi = b|βi (1.86)

and assuming that they commute, i. e. [Â, B̂] = 0, then they have common eigenvalues (see Homewor-
problems 1), i. e.,
Â|αβi = a|αβi, B̂|αβi = b|αβi (1.87)
Therefore one cannot choose as quantum numbers to label simultaneously the states the eigenvalues
of, e. g., L̂21 , L̂1z , L̂2x and L̂2z , since these two last operators do not commute with each other. But
1 Basic Quantum Mechanics concepts 19

there are many combinations one can choose. For instance, one can choose the eigenvalues of L̂21 , L̂1x ,
L̂22 , L̂2x . However, it is standard in Quantum Mechanics to choose as quantum numbers the eigenvalues
of the z-component of all angular momenta. Therefore the standard choose (which corresponds to all
existing Tables of angular momentum coefficients) is L̂21 , L̂1z , L̂22 , L̂2z or L̂21 , L̂22 , L̂2 , L̂z , i.e. the standard
eigenvectors used to label the angular momenta are

|l1 m1 l2 m2 i or |l1 l2 lmi (1.88)

and, therefore, the standard projectors are

X X
|l1 m1 l2 m2 ihl1 m1 l2 m2 | = Iˆ or |l1 l2 lmihl1 l2 lm| = Iˆ (1.89)
l1 m1 l2 m2 l1 l2 lm

One can write the vector in one representation in terms of the other representation, for instance
X
|l1 m1 l2 m2 i = |l1 l2 lmihl1 l2 lm|l1 m1 l2 m2 i, (1.90)
lm

and X
|l1 m1 l2 m2 i = hl1 m1 l2 m2 |lmi|l1 l2 lmi (1.91)
lm

The number hl1 m1 l2 m2 |lmi = hl1 l2 lm|l1 m1 l2 m2 i is real and is called Clebsch-Gordan coefficient. Due to
the orthonormality of the basis elements
X
|l1 l2 lmi = hl1 m1 l2 m2 |lmi|l1 m1 l2 m2 i (1.92)
m1 m2

If the Hamiltonian corresponding to the two-particle system is spherically symmetric then the eigen-
states of the Hamiltonian can be labeled by the eigenvalues of the angular momenta shown above.

Symmetry properties of the Clebsch-Gordan coefficient

The Clebsch-Gordan coefficient can best be written in terms of the 3-j symbol defined as

(−1)l1 −l2 +m
 
l1 l2 l
= √ hl1 m1 l2 m2 |lmi (1.93)
m1 m2 −m 2l + 1
with the properties that
     
l1 l2 l l2 l l1 l l1 l2
1. = =
m1 m2 m m2 m m1 m m1 m2
   
l1 l2 l l1 +l2 +l l2 l1 l
2. = (−1)
m1 m2 m m2 m1 m
   
l1 l2 l l1 l2 l
3. = (−1)l1 +l2 +l
−m1 −m2 −m m1 m2 m
4. m1 + m2 − m = 0

Short Algebraic Table of the 3-j Symbols

   1/2
j1 j2 j3 (J − 2j1 )!(J − 2j2 )!(J − 2j3 )!
= (−1)J/2
0 0 0 (J + 1)!
( 12 J)!
× (1.94)
( 12 J − j1 )!( 12 J − j2 )!( 12 J − j3 )!

where J = j1 + j2 + j3 is even.
1 Basic Quantum Mechanics concepts 20

   1/2
j j 0 1
= (−1)j−m (1.95)
m −m 0 2j + 1
1/2
j + 12 1 j − m + 12
  
j 1
1
2
1 = (−1)j−m− 2 (1.96)
m −m − 2 2 (2j + 2)(2j + 1)
   1/2
j+1 j 1 j−m−1 (j − m)(j − m + 1)
= (−1) (1.97)
m −m − 1 1 (2j + 3)(2j + 2)(2j + 1)
   1/2
j+1 j 1 j−m−1 2(j + m + 1)(j − m + 1)
= (−1) (1.98)
m −m 0 (2j + 3)(2j + 2)(2j + 1)
   1/2
j j 1 2(j − m)(j + m + 1)
= (−1)j−m (1.99)
m −m − 1 1 (2j + 2)(2j + 1)(2j)
 
j j 1 2m
= (−1)j−m (1.100)
m −m 0 [(2j + 2)(2j + 1)(2j)]
1/2

The orthogonality relations,

X  j1 j0
 
j2 j j1 j2
(2j + 1) = δjj 0 δmm0 . (1.101)
m1 m2 m m1 m2 m0
m1 m2
  
X j1 j2 j j1 j2 j
(2j + 1) = δm1 m01 δm2 m02 . (1.102)
m1 m2 m m01 m02 m
jm

6-j symbols
In the sum of three angular momenta one can choose the partition

J = j1 + j2 + j3 = J12 + j3 = j1 + J23 (1.103)

where
J12 = j1 + j2 , J23 = j2 + j3 (1.104)
One can write the basis vector in one representation in terms of the other representation as
X
|(j1 j2 )J12 j3 ; JM i = hj1 (j2 j3 )J23 ; J|(j1 j2 )J12 j3 ; Ji|j1 (j2 j3 )J23 ; JM i (1.105)
J23

The symmetry properties of the expansion coefficient can best be seen by introducing the 6-j symbol as

hj1 (j2 j3 )J23 ; J|(j1 j2 )J12 j3 ; Ji

 
p j1 j2 J12
= (−1)j1 +j2 +j3 +J (2J12 + 1)(2J23 + 1) (1.106)
j3 J J23

which is a real number (therefore it is the same for h(j1 j2 )J12 j3 ; J|j1 (j2 j3 )J23 ; Ji). The 6-j symbol does
not change if two columns are inter changed, for instance
   
j1 j2 J12 j1 J12 j2
= . (1.107)
j3 J J23 j3 J23 J

The 6-j symbol is also invariant if upper and lower arguments are interchanged in any two columns,
       
j1 j2 j3 j4 j5 j3 j1 j5 j6 j4 j2 j6
= = = (1.108)
j4 j5 j6 j1 j2 j6 j4 j2 j3 j1 j5 j3
.
The angular momentum triangular relation must be satisfied for (j1 , j2 , J12 ), (j1 , J, J23 ), (j3 , j2 , J23 )
and (j3 , J, J12 ). Thus, e.g.,  
1/2 1/2 0
=0 (1.109)
1/2 1/2 2
1 Basic Quantum Mechanics concepts 21

since 1/2 + 1/2 < 2.

In general, When j6 = 0 the expression for the 6-j symbol can be written as
 
j1 j2 j3 δj2 ,j4 δj1 ,j5
=p (−1)j1 +j2 +j3 {j1 , j2 , j3 }. (1.110)
j4 j5 0 (2j1 + 1)(2j2 + 1)

The function {} is equal to 1 when the triad satisfies the triangle conditions, and zero otherwise.
The 6-j symbols satisfy the orthogonality relation,
δj6 j60
  
X j1 j2 j3 j1 j2 j3
(2j3 + 1) 0 = {j1 , j5 , j6 }{j4 , j2 , j6 }. (1.111)
j4 j5 j6 j4 j5 j6 2j6 + 1
j3

9-j symbols
In the case of 4 angular momenta
J = j1 + j2 + j3 + j4 (1.112)
one can write, e.g.,
J = J12 + J34 = J13 + J24 (1.113)
where J12 = j1 + j2 , J34 = j3 + j4 , J13 = j1 + j3 and J24 = j2 + j4 .
One can thus write
X
|(j1 j3 )J13 (j2 j4 )J24 ; JM i = h(j1 j2 )J12 (j3 j4 )J34 ; J|(j1 j3 )J13 (j2 j4 )J24 ; JM i
J12 J34
×|(j1 j2 )J12 (j3 j4 )J34 ; Ji (1.114)

and the 9-j symbol is defined by,

h(j1 j2 )J12 (j3 j4 )J34 ; J|(j1 j3 )J13 (j2 j4 )J24 ; Ji

 
p  j1 j2 J12 
= (2J12 + 1)(2J34 + 1)(2J13 + 1)(2J24 + 1) j3 j4 J34 (1.115)
J13 J24 J
 

which is also a real number.

The symmetry properties of the 9-j symbols are
1. Any permutation of rows and columns does not change the 9-j symbol except the sign, which is plus
if the permutation is even and (−1)S , where S is the sum of all angular momenta, if the permutation
is odd.
For example, we have,
     
 j1 j2 j3   j4 j5 j6   j2 j1 j3 
j4 j5 j6 = (−1)S j1 j2 j3 = (−1)S j5 j4 j6 , (1.116)
j7 j8 j9 j7 j8 j9 j8 j7 j9
     

where
9
X
S= ji . (1.117)
i=1

2. The 9-j symbol does not change under a reflection about either diagonal.
     
 j1 j2 j3   j1 j4 j7   j9 j6 j3 
j4 j5 j6 = j2 j5 j8 = j8 j5 j2 (1.118)
j7 j8 j9 j3 j6 j9 j7 j4 j1
     

The 9-j symbols can be calculated as sums over triple-products of 6-j symbols where the summation
extends over all ”x” admitted by the triangle condition,
 
 j1 j2 j3  X    
2x j1 j4 j7 j2 j5 j8 j3 j6 j9
j4 j5 j6 = (−1) (2x + 1) . (1.119)
j8 j9 x j4 x j 6 x j 1 j2
j7 j8 j9
 
x
1 Basic Quantum Mechanics concepts 22

When j9 = 0 the 9-j symbol is proportional to a 6-j symbol as

 
 j1 j2 j3  δj3 ,j6 δj7 ,j8
 
j2 +j3 +j4 +j7 j1 j2 j3
j4 j5 j6 =p (−1) . (1.120)
(2j3 + 1)(2j7 + 1) j5 j4 j7
j7 j8 0
 

The 9-j symbols satisfy this orthogonality relation,

 j1 j2 j3   j1 j2 j30  δ 0 δ 0 {j j j }{j j j }{j j j }

  
X j3 j3 j6 j6 1 2 3 4 5 6 3 6 9
(2j7 + 1)(2j8 + 1) j4 j5 j6 j4 j5 j60 = . (1.121)
(2j 3 + 1)(2j 6 + 1)
j7 j8 j9 j7 j8 j9
  
j7 j8

1.6 Homework problems

Exercise 1:
Using the eigenvectors and eigenvalues
 of the Paulimatrixσz as a basis
 evaluate
 the eigenvalues and
0 1 0 −i 1 0
eigenvectors of σx and σy , where σx = , σy = and σz = .
1 0 i 0 0 −1

Exercise 2:
a) Why translating the system of coordinates by an amount +∆r the function Ψ(r) becomes Ψ(r −
∆r)? (Eq. (1.45) of Chapter 1).
b) Show that the parity operator is Hermitian.

Exercise 3:
Show that if the operators A and B commute, then they have common eigenvectors.

Exercise 4:
a) Show that |j1 j2 JM i = (−1)j1 +j2 −J |j2 j1 JM i
b) Evaluate the following Clebsh-Gordan coefficients:
i) hj + 1/2 j − 1/2 1/2 1/2|j ji,
ii) hj + 1/2 j + 1/2 1/2 − 1/2|j ji,
iii) hj − 1/2 j − 1/2 1/2 1/2|j ji and
iv) hj − 1/2 j + 1/2 1/2 − 1/2|j ji.

Exercise 5:
a) Which is the relation between m1 , m2 and m, and between j1 , j2 and j in the Clebsh-Gordon
coefficient CG = hj1 m1 j2 m2 |jmi.
b) Show that CG = hjmjm|JM i=0if 2j − Jis odd. Which is the value of J if j = 1/2?
9 2 1 1
1 72

c) Show that 5 2 = 0 and 2 1 1 = 0.
3
2 2 2
3 2 2
 
Chapter 2
Gamow states and the Berggren representation

One of the main difficulties when studying processes occurring in the continuum is that the relevant
physical processes are time-dependent. The system is sensitive to the initial conditions and can easily fall
into chaos. The Gamow model presents a way to study the continuum by time-independent formalisms
in the complex energy plane. The physical reason behind is that, when the system is trapped by a high
enough barrier, it can remain in a localized region of space for a considerable long time. In this case
the system can be treated as quasistationary. In the following we will present these time-independent
formalisms for the evolution of the system in the complex energy plane.

2.1 One-particle Hamiltonian in one dimension

The one-dimension Hamiltonian is.

 2 2 
~ d
− + V (x) Φn (x) = En Φn (x) (2.1)
2µ dx2
and we will consider the square well potential shown in Fig. 2.1. There can be discrete and continuum
states in this case, as seen in Fig. 2.1. To solve the eigenvalue problem given by Eq. (2.1) we notice that
there are two regions:
Region (1): 0 < x < a; V (x) = −V0
Region (2): x > a; V (x) = 0
There are also two possibilities for the energy: Continue (E = Ec > 0) and Bound (E = −Eb < 0)
 2 2 
~ d
Region (1): − − V0 Φ(1) (1)
n (x) = En Φn (x)
2µ dx2
 2 2 
~ d
Region (2): − Φ(2) (2)
n (x) = En Φn (x)
2µ dx2
with
2µ 2µ
q2 = (En + V0 ); k 2 = 2 En
~2 ~
the eigenvectors solution of the eigenvalue problem are
( (1)
Φn (x) = An eiqx + Bn e−iqx
Φ(2)
n (x) = Cn e
ikx
+ Dn e−ikx
To determine the constant An , Bn , Cn and Dn , the boundary conditions of continuity of density and
current have to be applied. In addition, since V (x) = ∞ for x 6 0, one has

Φ(1)
n (x = 0) = 0 =⇒ An + Bn = 0

I) Bound states
2µ 2µ
q2 = (V0 − Eb ) > 0; k 2 = − 2 Eb < 0
~2 ~
Notice that we assume Eb > 0 and, therefore, the energy of the bound state is −Eb .
r
2µEb
k = ±iχ; χ =
~2

(1) (2)
 Φn (a) = Φn (a)

d (1) d (2)
Φ (x) = Φ (x)
dx n dx n


x=a x=a
2 Gamow states and the Berggren representation 24

V(x)

Ec

0 a
x

Eb

-V0

Figure 2.1: Square well potential in one-dimension. The range of the potential is a and the depth is −V0 .
For x < 0 the potential is infinite and, therefore, the wave function vanishes at x = 0. Eb (Ec ) is the
energy of a bound (continuum) state.

(2)
An additional condition in Φn (x) = Cn e−χx + Dn eχx is that since eχx diverges as x → ∞ one has
to impose Dn = 0. Besides there is the normalization condition. With the constants thus evaluated
one obtains the possible energies as those for which the continuity relations are satisfied.
II) Continuum
2µ 2µ
q2 =(V0 + Ec ) > 0; k2 = Ec > 0
~2 ~2
assuming that the system is confined in the region
Z L
2
0 < x < L =⇒ |Φn (x)| dx = 1
0

Notice that all energies Ec > 0 are allowed in the continuum, but only a discrete number of energies
−Eb < 0 are allowed as bound states.
In these cases the projector becomes,
N
X Z
Iˆ = |nihn| + dα|αihα| (2.2)
n=1

where |ni (|αi) is a discrete (continuum) state.

The orthonormalization condition now reads,
N
X Z
δ(r − r 0 ) = Ψ∗n (r)Ψn (r 0 ) + dαΨ∗α (r)Ψα (r 0 ) (2.3)
n=1

2.2 Square barrier

This potential is defined as

V (x < 0) = ∞; V (0 < x < a) = 0; V (a < x < b) = V0 > 0; V (x > b) = 0
There are three regions.
Region I, 0 < x < a; VI (x) = 0
Schrödinger equation
~2 d2
HϕI (k, x) = − ϕI (k, x) = EϕI (k, x)
2m dx
2 Gamow states and the Berggren representation 25
√
i. e., (d2 /dx2 + k 2 )ϕI (k, x) = 0, where k = ± 2mE/~.
ϕI (k, 0) =0 =⇒ ϕI (k, x) = A sin(kx)

Region II, a < x < b; VII (x) = V0

~2 d2
HϕII (k, x) = (− + V0 )ϕII (k, x) = EϕII (k, x)
2m dx
i. e., (d2 /dx2 − κ2 )ϕII (κ, x) = 0, where κ2 = 2m(V0 − E)/~2 > 0.

ϕII (κ, x) = A1 eκx + A2 e−κx (2.4)

Region III, x > b; VIII (x) = 0

ϕIII (k, x) = Bo eikx + Bi e−ikx

Boundary conditions
1) At x=a
A sin(ka) = A1 eκa + A2 e−κa
Ak cos(ka) = κ(A1 eκa − A2 e−κa )
Which gives,
A1 = A∆1 /∆, A2 = A∆2 /∆, where
sin(ka) e−κa
∆1 = = −e−κa (κ sin(ka) + k cos(ka))
k cos(ka) −κe−κa
eκa sin(ka)
∆2 = = eκa (k cos(ka) − κ sin(ka))
κeκa k cos(ka)
eκa e−κa
∆= = −2κ
κeκa −κe−κa
2) At x=b
A1 eκb + A2 e−κb = Bo eikb + Bi e−ikb
κA1 eκb − κA2 e−κb = ikBo eikb − ikBi e−ikb
i. e.
Bo = ∆3 /∆0 , Bi = ∆4 /∆0 , where
A1 eκb + A2 e−κb e−ikb
∆3 =
κA1 eκb − κA2 e−κb −ike−ikb
eikb A1 eκb + A2 e−κb
∆4 =
ikeikb κA1 eκb − κA2 e−κb
eikb e−ikb
∆0 = = −2ik
ikeikb −ike−ikb

The ratio between the incoming and outgoing wave function amplitudes is
Bi ∆4
=
Bo ∆3

Outgoing boundary conditions

This corresponds to Bi = 0, i. e. ∆4 = 0

Phase shift
The wave function in region III can be written as,

ϕIII (k, x) = (Bo + Bi )cos(kx) + i(Bo − Bi )sin(kx)

2 Gamow states and the Berggren representation 26

calling B1 = Bo + Bi , B2 = i(Bo − Bi ), one gets

B1
ϕIII (k, x) = B2 (sin(kx) + cos(kx))
B2
and one defines the phase shift δ as
B1
tanδ =
B2
from where one gets,
ϕIII (k, x) = Csin(kx + δ) (2.5)
where C is a constant.
Notice that the difference between regions I and III is a phase shift in the wave function. This is the
reason of the name of the angle δ.
At the point x = b the continuity of the density and current can be expressed as the logarithmic
derivative of the wave function. This quantity, usually called β, is given by,

β = ϕ0II (x, b)/ϕII (x, b)

which is evaluated by using the expression (2.4). Continuity implies that it is also,

kcos(kx + δ)
β = ϕ0III (x, b)/ϕIII (x, b) = x=b
sin(kx + δ)

By writting sin(kx + δ) and cos(kx + δ) in terms of exp(i(kx + δ)) and exp(−i(kx + δ)) one gets,

β + ik
e2iδ = e−2ikb
β − ik
The quantity
S = e2iδ
is called S − matrix. If β is real then it is S ∗ = S −1 , that is the S-matrix is unitary.
The value of δ(E) when E cross the energy of a resonance changes by an angle π.

2.3 Gamow states

The wave function of a resonance with a peak at energy E0 and a width Γ can be factorized as
s
Γ/2
Φ(E, r) = Ψ(r), (2.6)
π [(E − E0 )2 + (Γ/2)2 ]
p
where Ψ(r) = πΓ/2 Φ(E0 , r). Through the Fourier transform, we obtain the time evolution of the
resonance Z ∞
Φ(t, r) = Φ(E, r)e−iEt/~ dE = Ψ(r)e−iẼt/~ , (2.7)
−∞

which gives us the resonance in the form of a stationary state, but with a complex energy
Γ
Ẽ = E0 − i . (2.8)
2
The probability of measuring the system at t is given by

|Φ(t, r)|2 = |Ψ(r)|2 e−Γt/~ . (2.9)

The half-life of the resonance can be obtained from

~ ln 2
T1/2 = . (2.10)
Γ
2 Gamow states and the Berggren representation 27

Figure 2.2: Radial function of a narrow resonance and a bound state. The solid and the dashed line
denote the real and imaginary part of the wave function of a narrow resonance respectively, while the
dotted line denotes the wave function of a bound state.

Therefore a time-dependent process has now been transformed into a stationary problem by going to
the complex energy plane. The complex solutions to the corresponding stationary Schrödinger equation
are called Gamow states. The wave numbers kn are a discrete set of complex values, which satisfy

~2 2
Ẽn = k . (2.11)
2µ n
They can be written as
kn = κn + iγn . (2.12)
The states can be classified into four classes, namely:
1. bound states, for which κn = 0 and γn > 0;
2. antibound states with κn = 0 and γn < 0;

3. decay resonant states with κn > 0 and γn < 0;

4. capture resonant states with κn < 0 and γn < 0.
Since the radial wave function has the form of w(En , r) ∼ eikn r , one can see that the wave function
of the resonant states will diverge at infinity. However a narrow resonance can still be treated stationary
since the wave function does not diverge at small distance. As shown in Fig. 2.2, the narrow resonance
has the same wave function as a bound state inside of the nucleus, but at a large radius far out from the
nuclear surface it starts to oscillate and diverge. Since we are interested in the processes occurring at the
scale of nuclear distances, we can eventually consider the narrow resonance as a bound state and solve
the time-independent Schrödinger equation approximately.
The antibound state has a pure negative imaginary wave number kn = −i|γn |, which gives the
corresponding energy real and negative, as the bound state. However unlike the bound state, whose wave
function diminishes exponentially at large distances, the antibound state diverges exponentially. As a
result, the antibound state can be physically meaningful only if it is sufficiently close to the threshold.
Therefore it may have a great influence on low-energy cross sections and nuclear spectroscopy.
We plotted in Fig. 2.4 the 1s1/2 antibound state. One sees that it extends in an increasing rate far
out from the nuclear surface, as expected in this halo nucleus (the standard value of the radius is here
1.2 × 111/3 =2.7 fm).
2 Gamow states and the Berggren representation 28

Figure 2.3: Radial function φ(r) corresponding to the single-particle neutron antibound state 0s1/2 at an
energy of -0.050 MeV. Taken from Ref. [1].

11
Figure 2.4: A schematic picture of the halo nucleus Li.
2 Gamow states and the Berggren representation 29

Figure 2.5: Integration contour L+ in the complex energy plane. The open circles denote the resonances
included in the sum of Eq. (2.14), while the solid circles are those excluded. The vertex (c, 0) corresponds
to the energy cutoff point c.

2.4 Berggren completeness relation

The eigenvectors of a Hamiltonian provide a representation projector which allows one to write
X Z ∞
δ(r − r0 ) = wn (r)wn (r0 ) + dEu(r, E)u(r0 , E), (2.13)
n 0

where wn (r) are the wave functions of the bound states and u(r, E) are the scattering states. The
integration contour is along the real energy axis, and all the energies and radial wave functions are real,
or more precisely, can be chosen to be real. Notice that only bound states and scattering states enter in
Eq. (2.13).
In order to include other Gamow states, Berggren extended the integration contour to the complex
energy plane. By using Cauchy theorem, one gets
X Z
δ(r − r0 ) = w̃n∗ (r)wn (r0 ) + dE ũ∗ (r, E)u(r0 , E), (2.14)
n L+

where wn (r) are now the wave functions for all the bound and antibound states plus the resonances lying
between the real energy axis and the integration contour L+ , as shown in Fig. 2.5. In principle the
contour should start at the origin (0, 0), and end at infinity (∞, 0). However, one usually cuts the basis
at a certain maximum energy as in any shell model, within which only a limited number of shells are
included.
The tilde over the wave function wn (r) denotes the mirror state of wn (r), namely k̃n = −kn∗ . With
this one can prove that w̃n∗ (r) = wn (r), and the same for the scattering state u(r, E). Therefore the
internal product in Eq. (2.14) is the wave function times itself, and not its complex conjugate. This
internal product is called the Berggren metric.
To make the representation useful in real calculations, one has to discretize the contour integral, i.e.
Z X
dEu(r, E)u(r0 , E) = hp u(r, Ep )u(r0 , Ep ), (2.15)
L+ p

where the quantities Ep are usually selected by the Gaussian method on the contour and hp are the
corresponding weights. Therefore the complete set of orthonormal vectors |ϕj i now includes all the
bound states, antibound states and resonances inside the contour, p i.e., hr|ϕn i = wn (r, En ), and also
the discretized scattering states on the contour, i.e., hr|ϕp i = hp up (r, Ep ). This is the Berggren
representation used in the complex shell model calculations.
2 Gamow states and the Berggren representation 30

2.5 Homework problems

Exercise 1: a) Find the eigenvalues and eigenvectors of the impulse operator in one dimension, i. e.
d
px ϕ(x) = kϕ(x), where px = ~i dx , within the range 0 ≤ x ≤ L and the boundary condition ϕ(0) = ϕ(L).
2
b) The same for px with ϕ(0) = ϕ(L) = 0.

Exercise 2: Write the unit operator (projector) corresponding to the Hilbert space spanned by the
representation given by the eigenfunctions ϕ(x) of the one-particle Hamiltonian

~2 d2
H=− + V (x),
2m dx2
where V (x) = ∞ for x = 0, V (x) = −V0 for 0 < x ≤ a and V (x) = 0 for x > a. Assume that the wave
function in the continuum is normalized to unity within the region 0 < x ≤ L.
 Bibliography

[1] Z.X. Xu, R.J. Liotta, C. Qi, T. Roger, P. Roussel-Chomaz, H. Savajols, R. Wyss, Nucl. Phys. A 850,
53-68 (2011).
Chapter 3
Nuclear Shell Model

Introduction to the shell model. Shell model representation. Shell model central potential. Shell closures
and the magic numbers. Particle excitations. Hole excitations.

3.1 Introduction to the shell model

The general Hamiltonian corresponding to the motion of the A=N+Z nucleons in a nucleus is
A A
X p2i X
H= + Vij (3.1)
i=1
2mi i<j=1

where mi is the mass of the ith nucleon (mi c2 ≈ 940 (938) MeV for neutrons (protons)) and Vij is
the nucleon-nucleon interaction. To find the eigenvectors and eigenvalues of this Hamiltonian exactly
is beyond of what is possible at present and even in the foreseeable future. The main reason why the
nuclear Hamiltonian cannot be treated is that one does not know how to evaluate the nucleon-nucleon
interaction starting from the underlying theory (QCD). It is not only this what is the problem. Even if
the interaction were known the task of treating the Hamiltonian (3.1) for medium and heavy nuclei would
be overwhelming. At a difference with other fields, like condensed matter, the number of nucleons in a
nucleus is never so large as to be possible to apply statistical concepts. However, there have been rather
successful attempts to perform just this task during the last couple of decades. In this calculations the
interaction is not the one provided by QCD, since this is unknown, but rather it is an effective interaction
that fulfills the symmetries required by QCD and also reproduces nucleon-nucleon data. Even with this
limitation one has not been able to treat nucleons with A > 16 so far. This procedure to solve the nuclear
many-body problem is called ”ab initio shell model” or ”no core shell model” because one starts just from
the nucleon-nucleon system (”ab initio”) and from there one adds nucleons up the A-particle system. No
core is present in this approach and therefore its name.
Another approach to obtain approximate solutions of the eigenvectors and eigenvalues of the nuclear
Hamiltonian (3.1) is based in the property that we have already discussed in the previous Chapter, i. e.
that double magic nuclei (i. e. nuclei with both N and Z magic numbers) are very stable. Some of the
points mentioned in that Chapter are very relevant here also and, therefore, we will insist on them. As
we have seen there, nuclear magnetic moments indicate that a nucleon moving outside a double-magic
core feels the interactions from the nucleons in the core as a whole. But this is not the only property
supporting the existence of magic numbers in nuclei. Other experimental evidences include:
1. The binding energies of magic-number nuclei is much larger than in the neighboring nuclei. Thus
larger energy is required to separate a single nucleon from magic nuclei.
2. The number of stable nuclei with magic values of Z or N is much larger than the corresponding number
in neighboring nuclei.
3. Naturally occurring isotopes with magic Z or N have greater relative abundances.
4. The first excited states in nuclei with magic numbers of neutrons or protons lie at higher energies than
the same states in neighboring nuclei.
5. Electric quadrupole moments of magic-number nuclei is zero as expected in closed shell nuclei, since
they should be spherically symmetric.
6. The energy of alpha or beta particles emitted by magic-number radioactive nuclei is larger than that
from other nuclei.
All this suggests that a double-magic nucleus is so stable that a particle moving outside this ”frozen
core” does not affect appreciably its internal structure. It is therefore reasonable to assume that the
frozen core induces a central field which only depends upon the distance r between the particle outside
3 Nuclear Shell Model 33

the core and the center of the core (notice that it does not depend on the direction of r, i. e. it is
spherically symmetric). This interaction, which we will call U (r), will be studied in detail below. Here
we will analyze its general features and its consequences on the nuclear many-body problem.
The nuclear Hamiltonian (3.1) can be written as
A A A
X p2 X X
H= ( i + U (ri )) + Vij − U (ri ) (3.2)
i=1
2mi i<j=1 i=1

Calling
A
X p2
H0 = ( i + U (ri )) (3.3)
i=1
2mi
and
A
X A
X
V = Vij − U (ri ) (3.4)
i<j=1 i=1

one gets,
H = H0 + V (3.5)
This is the shell model Hamiltonian.
We are assuming that for cases where there is only a particle outside a double-magic core the inter-
action U (r) describes the spectrum well and, therefore, in this situation it should be V = 0. When more
than one particle move outside the core, then V represents the interaction among these outside particles,
which are usually called ”valence particles”. It would be an impossible task to determine V from Eq.
(3.4). Instead one uses effective forms which reproduce as many experimental data as possible. One of
these effective forces is the pairing interaction mentioned in the previous Chapter but there are many
others, as we will see.
The formalism presented so far is based on the assumption that the bulk of the nuclear forces acting
upon the nucleons in a double-magic nucleus induces a frozen core. This is by far the major component
of the outcome of the nucleon-nucleon interaction and its consequence is reflected in the eigenvectors of
the central potential U . Therefore it is very suitable to use the eigenvectors of the Hamiltonian (3.3) to
describe the Hamiltonian (3.5), even when many valence particles are present, since its diagonalization
within such representation (as we did in Eq. (6) of Chapter 1) would need only a few of the infinite
members of the representation. This is an important point which is worthwhile to clarify farther. In
principle one can use as representation to diagonalize the shell model Hamiltonian the eigenvectors of
any other Hamiltonian, for instance the plane waves corresponding to free particles. But this would
require to include a huge number of vectors in the representation. Instead, very few eigenvectors of the
Hamiltonian (3.3) would be needed to describe accurately enough the nuclear properties. The dimensions
of the shell model Hamiltonian matrix is relatively small by using the shell model representation. Yet
the main advantage of this representation is not that the matrix dimensions are small but rather that it
provides a clear physical picture of the property that is analyzed. For instance, to describe a bound state
in terms of plane waves would require a huge number of basis states, all of them with approximately the
same value (and therefore very small, since the wave function should be normalized to one). Such wave
function contains no clue about the characteristics of the state under study. Instead, within the shell
model representation perhaps only one state may be enough, and the single-particle components in this
state depict clearly the nature of the state. This is the essence and reason of the success of the Shell
Model, which we will analyze in detail throughout this Course.

3.2 Shell model representation

The shell model representation consists of the eigenvectors of the Hamiltonian (3.3) where each of the A
nucleons move under the influence of the central potential U . Calling H0 (i) the Hamiltonian corresponding
to the ith particle, one has
XA
H0 = H0 (i) (3.6)
i=1
3 Nuclear Shell Model 34

the eigenvectors |ϕni > and eigenvalues ni of H0 (i) satisfy

p2i
H0 (i) < ri |ϕn >= ( + U(ri )) < ri |ϕni >= ni < ri |ϕn > (3.7)
2mi
Since the Hamiltonian H0 , with eigenvalues given by

H0 |Ψα >= Eα |Ψα > , (3.8)

is a sum of the Hamiltonians H0 (i) and the degrees of freedom of different particles are independent of
each other, the states |Ψα > can be written as,

< n1 r1 , n2 r2 ....nA rA |Ψα >= Ψα (n1 r1 , n2 r2 , ...nA rA ) = ϕn1 (r1 )ϕn2 (r2 )...ϕnA (rA ) (3.9)

and the eigenvalues are

Eα = n1 + n2 + ... + nA (3.10)
From < ϕi |ϕj >= δij , it follows that < Ψα |Ψβ > = δαβ .
Assuming that one chooses N of the infinite number of single-particle states, the representation to be
formed with the functions (3.9) consists of all possible combinations of the A particles in the different
states |ni=1,2,...,N >. However one has still to consider the Pauli principle, which requires that when two
particles change position the wave function changes sign. This is called antisymmetric wave function.
Such a function can be built as a linear combination of the functions (3.9). For instance, assuming A =
2 a particular antisymmetric wave function would be

Ψ(a)
α (n1 r1 , n2 r2 ) = Nα (ϕn1 (r1 )ϕn2 (r2 ) − ϕn1 (r2 )ϕn2 (r1 )) (3.11)
(a) (a)
which has the property that Ψα (n1 r1 , n2 r2 ) = −Ψα (n1 r2 , n2 r1 ). The constant
√ Nα has to be included
due to the normalization of the wave function, which requires that Nα = 1/ 2. Since ϕn1 (r2 )ϕn2 (r1 ) =
(a)
ϕn2 (r1 )ϕn1 (r2 ) one sees that two particles cannot occupy the same state since, in that case, Ψα (n1 r1 , n1 r2 ) =
0. In general the antisymmetric wave function satisfies,

Ψ(a) (a)
α (n1 r1 , n2 r2 , ..., np rp , ..., nq rq , ..., nA rA ) = −Ψα (n1 r1 , n2 r2 , ..., np rq , ..., nq rp , ..., nA rA ) (3.12)

We will analyze this general case later, but here it is important to stress that the many-body represen-
(a)
tation is the set of states {Ψα (n1 r1 , n2 r2 , ..., np rp , ..., nq rq , ..., nA rA )} that can be formed through the
combination of the A particles placed in the different states ni=1,2,...,N .
The property that two particles cannot be in the same single-particle state implies that the lowest
state of the non-interacting A-particle system consists of a particle occupying the lowest single-particle
state, the second particle occupying the second lowest state and so on. Such a system of non-interacting
particles fulfilling the Pauli principle is called ”a Fermi gas”.
The important conclusion of this discussion is that the potential U should provide a set of single-
particle states which, placing the particles in the order required by the Fermi gas, explains the appearance
of magic numbers. To find such a potential is the task that we will confront now.

3.3 Shell model central potential

We have solved the Schrödinger equation corresponding to a one-dimension square well potential in the
previous Chapter. We will now proceed to the more realistic case of three dimensions following the
historical developments in this subject. In the beginning the spin of the nucleon was neglected and
therefore we will start in spherical coordinates with a Hamiltonian corresponding to a spinless particle
moving in the central field U (r), i. e.

p2 l̂2 ~2 ∂
 
∂
H= + U (r) = 2
− r2 + U (r) (3.13)
2m 2mr 2mr2 ∂r ∂r

One sees that the angular coordinates (θϕ) are included in the orbital angular momentum operator l̂2
only while the radial coordinate r is in the rest of the Hamiltonian. Since there is no term which mixes
3 Nuclear Shell Model 35

them, the eigenfunction of this Hamiltonian is a product of radial and angular coordinates. As we have
already seen, with
l̂2 Ylm (θϕ) = ~2 l(l + 1)Ylm (θϕ) (3.14)
one can write the total eigenfunction as

Ψnlm (r) = Rnl (r)Ylm (θϕ), (3.15)

It is convenient to replace the radial eigenfunction Rnl (r) by the function unl (r) defined as

Rnl (r) = unl (r)/r (3.16)

and the eigenvalue problem acquires the simpler, one dimensional, form,

~2 d2 l(l + 1)~2
 
− unl (r) + + U (r) unl (r) = Enl unl (r) (3.17)
2µ dr2 2µr2

The first problem that had to be confronted by the pioneers in the study of the structure of nuclei
was to determine the form of the central potential. For this purpose it was taken into account that the
nuclear force is very strong and of a very short range. Besides, the binding energy of the particle bound
to the core is strong, of the order of several MeV. Therefore the mean field itself should be deep and
of a short range. A convenient potential to describe such states which, as seen in Fig. 3.1, lie far from
the continuum threshold, is an harmonic oscillator potential. A more realistic one is the Woods-Saxon
potential UW S given by,
U0
UW S (r) = − (r−R
(3.18)
1+e 0 )/a

where U0 is a positive constant such that UW S (r) is attractive, R0 is the radius of the nucleus, which is
taken to be R0 ≈ 1.2A1/3 fm and a is the diffuseness, which usually has a value around 0.5-0.6 fm.
In the next Chapter we will introduce the Hartree-Fock potential which is a nuclear mean field derived
microscopically i. e. from the degrees of freedom of the nucleons that constitute the nucleus and their
interactions.
The use of the harmonic oscillator potential has the advantage of providing eigenvectors, i. e. a
representation, which is very convenient in many-body studies. One thus describes bound states lying
far from the continuum threshold by using a Harmonic oscillator potential of the form U (r) = mω 2 r2 /2

~2 d2 l(l + 1)~2
 
1 2 2
− u nl (r) + + mω r unl (r) = Enl unl (r) (3.19)
2m dr2 2mr2 2

The eigenvalues corresponding to this potential are

3
Enl = (2n + l + )~ω (3.20)
2
where n ≥ 0 is and integer. One sees that the energy depends only upon the number N , which is called
”principal quantum number”, given by
N = 2n + l (3.21)
which, like n, is also integer and N ≥ 0. The energy becomes
3
EN = (N + )~ω (3.22)
2
Since ~ω is a free parameter that defines the Harmonic oscillator potential, it can be chosen large enough
so that there are in the spectrum a series of bands, each band labelled by the number N . The gap
between bands was found to be related to the mass number A by the relation,

~ω = 45A−1/3 − 25A−2/3 , (3.23)

thus allowing to determine the frequency ω in different regions of the nuclear chart.
It is important to notice that the lowest (ground) state, i. e. the one corresponding to N = 0, is
not at zero energy but rather at EN =0 = (3/2)~ω. This is a Quantum Mechanics effect. It is as if there
is a motion due to the uncertainties inherent to Quantum Mechanics that produces an energy even if,
classically, there is no real motion. This energy is therefore called ”zero-point energy”. It shows that in
3 Nuclear Shell Model 36

10

V(r) (MeV) -10

-20 Woods-Saxon
Harmonic oscillator
-30

-40

-50

-60

-70
0 1 2 3 4 5 6 7 8 9 10
r

Figure 3.1: Potential describing bound single-particle states. The nuclear central field is well approxi-
mated by the Woods-Saxon potential (3.18). In the region where the single-particle states lie the nuclear
(Woods-Saxon) potential can be approximated by an Harmonic oscillator potential. Close to the con-
tinuum threshold the nuclear potential vanishes and the centrifugal plus Coulomb potentials become
dominant.

Quantum Mechanics (particularly in Field Theories) the vacuum is not a real vacuum. Everywhere there
is the zero-point energy associated to the ground states of all fields.
The spectrum of the Harmonic oscillator is given by Eq. (3.22). Each value of N implies several
values of (n, l), following Eq. (3.21). The corresponding level sequence is as shown in Table 3.1. One
sees in this Table that between one value of N and the next there is an energy gap ~ω. Therefore the set
of (n, l) values corresponding to a particular value of N would be clustered together in bands separated
by an energy ~ω. In the Table the levels are labelled by (n, l), where the states l = 0, 1, 2, 3, 4, .. are
traditionally called s, p, d, f, g....

Table 3.1: Energy levels corresponding to an Harmonic oscillator potential of frequency ω. The states
are labelled by N = 2n + l. The energies Enl are in units of ~ω. Dl is the degeneracy of the state (n, l).
P
N n l Enl |nli parity Dl = 2(2l + 1) Dl
0 0 0 3/2 0s + 2 2
1 0 1 1+3/2 0p - 6 8
0 2 2+3/2 0d + 10
2 20
1 0 2+3/2 1s + 2
0 3 3+3/2 0f - 14
3 40
1 1 3+3/2 1p - 6
0 4 4+3/2 0g + 18
4 1 2 4+3/2 1d + 10 70
2 0 4+3/2 2s + 2

Since there is no term in the Hamiltonian that mixes the l and s angular momenta, the levels for the
same N are degenerate, as seen in the level scheme labelled by (N + 3/2)~ω in Fig. 3.2. It is important
to notice that the notation in this Figure is different than the one in the Table. In the Table we have
used the Harmonic oscillator quantum values for the labels. Therefore the lowest state correspond to
n = 0. Besides providing the energy, the number n is the number of nodes of the function unl (r) in Eq.
(3.16) (the number of nodes is the number of times that unl (r) vanishes). This function has always a
node since unl (r = 0) = 0 for all l. Therefore some shell model practitioners label the lowest state with
3 Nuclear Shell Model 37

the value n = 1, as in the Figure. We will use throughout this course the convention that the lowest state
corresponds to n = 0.
As with the magnetic moments discussed in the previous Chapter, this different conventions for the
value of n is unfortunate, but it should not produce any confusion.

3.4 Shell closures and the magic numbers

We are now in a position to give an explanation to the shell closures that give rise to the magic numbers,
i. e. the numbers 2, 8, 20, 28, 50, 82 and 126. As we have discussed above, experimental data show that
nuclei with a magic number of neutrons or protons have properties that are very different than other
nuclei. In particular, if N and Z are both magic the nucleus is very stable and can be considered a frozen
core.
One has to consider the spherical symmetry of the potential in order to count the number of particles
that can be placed in each state satisfying the Pauli principle. That symmetry implies that the energy
do not depend upon the projection lz of the angular momentum, i. e. upon the quantum number m.
Therefore in a given state (nl) one can place 2l+1 particles, each in one of the 2l+1 values of m (remember
that −l 6 m 6 l). In addition each particle has its own intrinsic spin s = 1/2, which gives an additional
degeneracy 2s + 1 = 2. Therefore the total degeneracy of the state Nl = 2n + l is Dl = 2(2l + 1). As seen
in Table (3.1) this procedure reproduces the lowest magic numbers, i. e. 2, 8 and 20, but not the others.
Besides, it was found that experimentally different values of l have different energies. In an attempt to
explain this feature an attractive term proportional to Dl2 , with D a negative number, was included.
This contributes to the energy with the value Dl(l + 1), as seen in the corresponding level scheme of Fig.
3.2. But still the sequence of the levels was not well explained.
In 1949 Goeppert-Mayer and, independently, Haxel, Jensen and Suess proposed the existence of an
strong spin-orbit potential that could account for the observed magic numbers. For this Goeppert-Mayer
and Jensen got the 1963 Physics Nobel Prize (it was the last time so far that a lady, Maria Goeppert-
Mayer, was awarded the Nobel Prize in Physics. Incidentally, the other lady who is a physics Nobel
laureate, Marie Curie, was also a nuclear physicist).
The spin-orbit potential proposed by Goepper-Mayer and Jensen et. al. has the form, Vso (r)l · s,
where l is the orbital angular momentum and s is the intrinsic spin of the nucleon. The shell model
Hamiltonian acquires the form,
p2 1
H= + mω 2 r2 + Vso (r) l · s (3.24)
2m 2
The orbital and spin angular momenta must be coupled to a total angular momentum j = l + s. As we
have seen in Chapter 1 the eigenstates of the spin-orbit term are determined by the quantum number
j = |l ± 1/2| and by the quantum number m associated with the third component of j. Therefore the
corresponding eigenvectors are {|nlsjmi} and with

j 2 = l2 + s2 + 2 l · s (3.25)

l · s = (j 2 − l2 − s2 )/2 (3.26)
one gets,
~2  
l · s|nlsjmi = j(j + 1) − l(l + 1) − s(s + 1) |nlsjmi (3.27)
2
In r-representation it is,
 
hr|nlsjmi = Rnlj (r) Yl (θϕ)χ1/2 jm , Rnlj (r) = unlj (r)/r (3.28)

and the Schrödinger equation becomes

(
~2 d2 l(l + 1)~2 1
− 2
unl (r) + 2
+ mω 2 r2 +
2m dr 2mr 2
)
~2  
+ j(j + 1) − l(l + 1) − 3/4 Vso (r) unl (r) = Enl unl (r). (3.29)
2
3 Nuclear Shell Model 38

Figure 3.2: In the left is the energy spectrum corresponding to a Harmonic oscillator potential of frequency
ω. In the middle is the same plus a term Dl2 , with D negative. The spectrum to the rigth corresponds
to the shell model Hamiltonian (3.5), which follows very well the tendencies of the experimental data.
3 Nuclear Shell Model 39

Since |l − 1/2| 6 j 6 l + 1/2 and j = l ± 1/2 one gets


  l; j = l + 1/2
j(j + 1) − l(l + 1) − 3/4 Vso (r) = Vso (r) (3.30)
−l − 1; j = l − 1/2

and one understands the experimental level scheme by assuming an attractive spin-orbit interaction, for
instance a constant term Vso (r) = −V0 . With this term one gets the level scheme labeled by Cl.s in Fig.
3.2.
One sees from Eq. (3.30) that the spin-orbit interaction is strongest for the state with the larger value
of the spin j and, due to the attractive character of the interaction, the total energy of this state lies
below the corresponding energy for the lower spin. This has as a consequence that the state in a band
with highest spin is so strongly affected by the spin-orbit interaction that they become mixed with states
of the band below. This can be observed in Fig. 3.2, where for l > 2 the states with highest spin, for
instance 1f7/2 , 1g9/2 , etc, crosses the number N below, mixing with the states N − 1. These are called
”intruder” states, and there appearance is due to the spin-orbit interaction. The parity of the intruder
state is (−1)N , i. e. it has different parity than the ones belonging to the band N − 1.
The states in the shell model level schemes, i. e. the one at the right in Figure 3.2, carry as quantum
number the total spin j and therefore, again due to spherical symmetry, each level has a degeneracy of
2j + 1. As seen in the Figure, one can then explain very well all magic numbers. One also sees that the
intruder states play a fundamental role. In particular, the intruder state 1f7/2 contains the eight nucleons
necessary to explain the magic number 28.
The spin orbit potential can be shown to have the form
1 dUW S (r)
Vso (r) = −vso . (3.31)
r dr
where vso is a positive number and UW S (r) is the Woods-Saxon potential (3.18). Due to the form of
this potential the spin orbit potential above is peaked in the surface of the nucleus and, therefore, its
influence is mainly felt just in that surface.
The presence of the spin-orbit potential was found to be a result of relativistic effects.

Parity

Since we assume that the nucleon-core potential is spherically symmetric the quantum numbers l.j, m
are conserved (as discussed in the previous Chapter). This potential will also be assumed to be invariant
under reflections, and therefore the parity π of the state |nljm > will also be conserved. To find the
value of the parity one has to analyze the Spherical Harmonics Ylml (θϕ) which, for ml > 0, is given by
r s
2l + 1 (l − ml )!
Ylml (θϕ) = (−1)ml eiml ϕ Plml (cos θ) (3.32)
4π (l + ml )!

where
(−1)ml (l + ml )! 2 d
l−ml
Plml (ξ) = (1 − ξ ) (ξ 2 − 1)l (3.33)
2l l! (l − ml )! dξ l−ml
For ml < 0 it is,
∗
Ylml (θϕ) = (−1)ml Yl−m l
(θϕ) (3.34)
The parity transformation corresponds to r −→ −r, that is (r, θ, ϕ) −→ (r, π − θ, ϕ + π), and since
for spherically symmetric potentials the value of r is the same for all values of the angles, only the
transformation of the Spherical Harmonics has to be considered. Therefore one finds,

Ψnljm (−r) −→ (−1)l Ψnljm (r) (3.35)

The spin-orbit part of the wave function is as above independently of the spherically invariant poten-
tial, but the radial part has to be studied separately for each potential one has to deal with.
We have seen that the parity associated to this wave function is (−1)l and since N = 2n + l the parity
also is (−1)N . That is, all the states in a band corresponding to the quantum number N has the same
parity.
Analytic expressions for the first few orthonormalized spherical harmonics in the Condon-Shortley
phase convention,
3 Nuclear Shell Model 40

r
1 1
Y00 (θ, ϕ) = (3.36)
2 π
r
1 3
Y1−1 (θ, ϕ) = sin θ e−iϕ (3.37)
2 2π
r
1 3
Y10 (θ, ϕ) = cos θ (3.38)
2 π
r
−1 3
Y11 (θ, ϕ) = sin θ eiϕ (3.39)
2 2π
r
1 15
Y2−2 (θ, ϕ) = sin2 θ e−2iϕ (3.40)
4 2π
r
1 15
2−1 (θ, ϕ) = sin θ cos θ e−iϕ (3.41)
2 2π
r
1 5
Y20 (θ, ϕ) = (3 cos2 θ − 1) (3.42)
4 π
r
−1 15
Y21 (θ, ϕ) = sin θ cos θ eiϕ (3.43)
2 2π
r
1 15
Y22 (θ, ϕ) = sin2 θ e2iϕ (3.44)
4 2π

Particle excitations

We have seen that a nucleus A Zm XNm , where Nm and Zm are magic numbers, acts as a frozen core. The
total angular momentum of this core is zero and a nucleon moving outside it feels the interaction of the
nucleons inside the core as a whole. The core induces a central field which generates the shell model
set of single-particle states, that is the shell model representation of Fig. 3.2. The nucleons in the core
form a Fermi gas such that each one of these nucleons occupy a level of the shell model representation.
Since the core is in the lowest excitation the nucleons occupy the lowest states. For instance, the nucleus
40
20 Ca20 (Calcium 40) is a core, since 20 is a magic number. In terms of the shell model the 20 neutrons as
well as the 20 protons occupy the states 1s1/2 , 1p3/2 , 1p1/2 , 1d5/2 , 2s1/2 and 1d3/2 . In the same fashion,
adding a nucleon, for instance a neutron, one would get the spectrum of 41 20 Ca21 . The lowest (ground)
−
state of this Ca isotope is, according to the shell model (Fig. 3.2), f7/2 , that is 7/2 , as indeed it is
57 −
experimentally. The ground state of 28 N i29 (Nickel 57) should be 3/2 , the first excited state should be
− −
5/2 and the second excited state 1/2 , as also it is experimentally.
The nuclear experimental spectra are found in many publications, but a convenient way of getting
them is through Internet. In the site
http : //www.nndc.bnl.gov/nudat2/
one finds all the available nuclear experimental spectra.
It is important to stress that the analysis we have carried out so far is based on the assumption that
the nucleus consists of freely moving nucleons in the shell model potential fulfilling the Pauli principle.
Considering that the nucleus is a bunch of strongly interacting nucleons packed in an extremely dense
environment it may seem nearly as a wonder that the independent particle shell model works so well.
There are arguments which allow one to justify this feature, but the importance of the single-particle shell
model lies not so much in its predicting power but rather in that it provides an excellent representation
to describe many-nucleon states, as we will see.
In order to complete the representation one has to know the energies of the single-particle states.
These can be taken from experiment, or from calculations that fit the experimental data as closely as
possible.
The single-particle energies ni are the eigenvalues of the Hamiltonian H0 (i), Eq. (3.7). They are
depicted in Fig. 3.3, where the origin of energies is the continuum threshold. That implies that all
values of  in the Figure are negative, as it should be for bound states. At positive energy there is the
continuum, which does not have any quantized state and, therefore, all energies are allowed. These are
all the possible kinetic energies of the particle at infinite, where the nuclear potential vanishes. One sees
3 Nuclear Shell Model 41

0
εh 1 εh 2 εp 1 εp 2

F L
V (r)

0 2 4 6 8 1 0
r
Figure 3.3: Single particle states in the shell model potential. The energies  are measured from the
continuum threshold. The Fermi level is indicated as FL.

from the Figure that ni is the energy necessary to bring a particle in the bound state ni to a rest state
(zero kinetic energy) at infinity. This is called ”separation energy”, which thus is the same as − ni .
We have seen that the A nucleons in a double magic core occupy the lowest shell model single-particle
states. In Fig. 3.3 these states are labeled by hi , where i = 1, 2..., A, and the highest one is called ”Fermi
level”. From the Fermi level and below all states are occupied. Above the Fermi level all states are
empty, i. e. unoccupied. They are labeled pk in the Figure, where k = 1, 2, ..., N and N is the number of
unoccupied states. Therefore, and since we assume that the core is frozen and neutral, the nucleus with
A + 1 nucleons has to be in one of the states pk .
According to Eq. (3.10) the energy of the core is
A
X
Ecore = hi (3.45)
i=1

and the energy of A + 1 nucleus in the state pk is

EA+1 (pk ) = Ecore + pk (3.46)

from where one gets,

pk = EA+1 (pk ) − Ecore (3.47)
Since all single-particle energies are negative one finds that

EA+1 (pk ) < Ecore (3.48)

The experimental values of EA are tabulated in many publications. A convenient way of getting them
is also through Internet. In the site
http://ie.lbl.gov/toi2003/MassSearch.asp
one finds the available binding energies B(A) of all nuclei in the ground state. The binding energy is
defined as B(A) = -EA , i. e. B(A) is a positive quantity such that BA+1 (pk ) > Bcore , according to Eq.
(3.48).
3 Nuclear Shell Model 42

+ +
3 .0 1 0 7 /2 , 9 /2
-
3 .0 0 7 3 /2
-
2 .5 7 7 7 /2 -
2 .4 4 3 5 /2

-
1 .1 1 3 1 /2
-
0 .7 6 9 5 /2

-
0 .0 3 /2
5 7
N i

Figure 3.4: Experimental spectrum of the nucleus 57 Ni29 taken from the Internet site mentioned in the
text. Energies are in MeV. The level at 3.010 MeV has not been completely determined yet. It can be
7/2+ as well as 9/2+ .

To obtain the single-particle energies from experiment one uses Eq. (3.47). We will show this, as well
as the limitations of the independent single-particle model, in an illuminating example.
As we have seen above, the neutron single-particle states between 28 and 50 can be extracted from
the spectrum of 57 Ni29 . The experimental spectrum in Fig. 3.4 shows that the ground state is 3/2− , i.
e. the state 1p3/2 (remember that we use the convention that for the lowest state it is n = 0). The first
−
excited state lies at 0.769 MeV and is 5/2 , i. e. 0f5/2 . The second excited state lies at 1.113 MeV and
−
is 1/2 , i. e. 1p1/2 . After this there are a number of states which do not correspond to the ones in Fig.
+
3.2. Only at the rather high energy of 3.010 MeV appears a state that may be 9/2 . This indicates that
after the state 1p1/2 the energy is high enough to disturb the core, probably by exciting single-particle
+
states from below to above the Fermi level. But one can still assume that the states 9/2 at 3.010 MeV
is the 0g9/2 level of Fig. 3.2.
The experimental energies in the spectrum are given relative to the ground state of the nucleus, as
seen in Fig. 3.4. To obtain the corresponding values of  one has to evaluate the binding energies B(Z, N ).
In the Internet site mentioned above are given the values of B(Z, N )/A. For our case we need the binding
energies of 56 Ni (the core) and of 57 Ni (see Eq. (3.47)), It is B(28, 28)/56 = 8.643 MeV and B(28, 29)/57
= 8.671 MeV, i. e. 1p3/2 = - 10.239 MeV. This is the lowest state. It corresponds to the state labelled
p1 in Fig. 8.7. The state 0f5/2 , at 0.769 MeV, corresponds to p2 (is less negative). Therefore one gets
0f5/2 = (- 10.239 + 0.769) MeV = - 9.470 MeV. In the same fashion it is 1p1/2 = -9.126 MeV and 0g9/2
= - 7.229 MeV.
The evaluation of the single-particle energies follows this procedure in all cases.

Hole excitations

We will now analyze the spectrum of a nucleus with a particle less than a magic number. According to
the shell model, the structure of such a nucleus is determined by filling the single-particle states as if the
system were a Fermi gas. This we did above with the core and with the core plus one-particle nucleus.
In the present case all the A-1 particles will occupy the levels below the Fermi level, i. e. the levels
called hi in Fig. 3.1. As seen in Fig. 3.5, the resulting spectrum is that of the core (all levels below the
Fermi level are occupied) except the level, called h1 in the Figure, where the last particle was not placed.
Therefore this excitation is called ”hole” since it is like a hole in the spectrum of the core. The energy of
3 Nuclear Shell Model 43

Figure 3.5: Excitations hi in the (A-1)-particle nucleus. It looks like at the level h1 there is a hole
in the completely filled states of the core. Therefore the states below the Fermi level are called ”hole
excitations”. A hole at the more deeply bound level h2 induces a higher excitation than the one at h1 .

the nucleus referred to the core is now (see Eq. (3.47))

EA−1 (h1 ) − Ecore = −h1 (3.49)

which is a positive number. This implies that a state which is more bound than h1 , for instance the one
labeled h2 in Fig. 3.5, lies at higher energy than h1 since |h2 | > |h1 |.
We will again clarify this point with an example, which is the spectrum of the nucleus 131 Sn, with
Z=50 and N=81. That is, the single-particle excitations of this nucleus are holes in the neutron core
below N=82. Therefore the ground state should be the first hole state, the first excited state the second
hole state and so on. And indeed we see that the level sequence of the experimental spectrum of Fig.
3.6 is the one predicted by the independent-particle shell model in Fig. 3.2. That is, the shell model
(experimental) levels are 1d3/2 (ground state), 0h11/2 (0+x MeV), i. e. experimentally has not been
possible yet to disentangle this from the ground state. Notice also that this is an intruder state. After
these states follow the levels 2s1/2 (0.332 MeV), 1d5/2 (1.655 MeV) and 0g7/2 (2.434 MeV). In this case
there is a good agreement between the experimental and the shell model sequence of levels, since only the
last two states are in a reverse sequence (the first two states are nearly degenerate). But in general one
expects that the levels predicted by the shell model should be there, but not with the exact sequence.
The ground state single-particle energy is, according to Eq. (3.49), h1 = Ecore − EA−1 (h1 ), where h1
is the hole state 1d3/2 and the core is 132 Sn(gs) (gs means ground state). The binding energies of interest
are B(50, 81)/131 = 8.363 MeV and B(50, 82)/132 = 8.355 MeV. Therefore it is Ecore = -B(50, 82) =
-1102.860 Mev and EA−1 (h1 ) = -B(50, 81) = -1095.553 MeV, from where one gets 1d3/2 = -7.307 MeV.
For the other states it is 0h11/2 = -7.307 MeV also, 2s1/2 = (-7.307 - 0.332) MeV = -7.639 MeV, 1d5/2
= -8.962 MeV and 0g7/2 = -9.741 MeV.

3.5 Parameterization of the Woods-Saxon Potential

arXiv:0709.3525
A number of parameterizations of the Woods-Saxon potential have been published, created with
different objectives and relevant to different nuclear mass regions. Most commonly used is the so-called
“Universal” parameterization [3] , which was adjusted to reproduce the single-particle binding energies
3 Nuclear Shell Model 44

+
2 .4 3 4 (7 /2 )

+
1 .6 5 5 (5 /2 )

+
0 .3 3 1 (1 /2 )
-
0 + x (1 1 /2 )
+
0 (3 /2 )
1 3 1
S n

Figure 3.6: Experimental spectrum of the nucleus 131

50 Sn81 . The levels are in parenthesis, meaning that
they are not completely determined yet. The level 11/2− is at 0.0+x MeV, implying that it is only
slightly above the 3/2+ ground state.

of proton and neutron orbitals around the doubly-magic nucleus 208 Pb and correct ground state spins for
nuclei of masses around A=180, but claimed to be applicable to lighter mass regions as well. Characteristic
for the “Universal” parameterization is the choice of different radii for the proton- and neutron potentials.
It has been pointed out, that this parameterization has shortcomings with respect to lighter nuclei and
that it predicts charge radii inconsistent with experiment. [4].
Basis for this investigation are the experimental single particle spectra around the doubly magic nuclei
16
O, 40 Ca, 48 Ca, 56 Ni, 132 Sn and 208 Pb. These systems were chosen, because they are expected to show
the most pure experimental manifestation of single-particle excitations.
Woods and Saxon [2] suggested to model the nuclear mean field i.e. the nucleon-core interaction with
a spherically symmetric potential that has a Fermi-function form
  −1
r−R
f (r, R, a) = 1 + exp , (3.50)
a
where the size R and diffuseness of the surface a are fixed parameters of the same units of length as r.
The total nuclear potential is defined as

V (r) = −V f (r, R, a), (3.51)

where V represents total strength and the minus sign is introduced to represent the attractive nature of
the interaction.
The electromagnetic force is a second part contributing to the proton-core interaction. This repulsive
potential is fully determined with the assumption of a given nuclear charge distribution ρ(r). The solution
of the corresponding electrostatics problem gives
 Z r Z ∞ 
1 02 0 0 0 0 0
Vc (r) = 4πe r ρ(r )dr + r ρ(r )dr . (3.52)
r 0 r

In the spirit of the Woods-Saxon parameterization it is often assumed that the nuclear charge distri-
bution is proportional in shape to the same function (3.50) ρ(r) ∼ f (r, Rc , ac ), where the coefficient of
proportionality must be determined from the normalization of density to the total nuclear charge. The
3 Nuclear Shell Model 45

integration in Eq. (3.52) along with a normalization of density must be done numerically, which is of-
ten too time consuming. The influence of surface terms on the strength of the Coulomb interaction is,
however, weak. We have numerically tested, that the diffuseness of the charge distribution can be set to
zero within the precision of the fit discussed below. Furthermore, for the same reason we have assumed
Rc = R which removes an extra unnecessary parameter that has little influence on the outcome. Except
for special cases [?] these assumptions are typical in other Woods-Saxon parameterizations. Through this
paper we adopt the following form of the Coulomb potential

(3R2 − r2 )/(2R3 ), r ≤ R,

0 2
Vc (r) = Z e (3.53)
1/r, r > R,

which as a result of the above assumptions corresponds to a uniformly charged sphere of radius R, which
can be treated analytically.
Thus, the total effective Hamiltonian becomes
p2
 
1 ∂
H= + V (r) + Vc (r) + 2 Ṽ (r) l · s, (3.54)
2µ 2µ r ∂r

where – unlike for the Coulomb field – the potential Ṽ (r) is not equal to the original potential V (r) and
may have a different form factor [?]. Therefore, the form factor of Ṽ (r) is another assumption that goes
into construction of the Woods-Saxon Hamiltonian

Ṽ (r) = Ṽ f (r, RSO , aSO ). (3.55)

Here, RSO and aSO stand for the radius and the diffuseness of the spin-orbit term.
The parameters of the potential change as one goes over the nuclear chart. The dependence of the
above 7 parameters on the number of protons and neutrons in the core-nucleon problem defines the
structure of parameterization. In the following, we will describe our choice of structure parameterization,
which we call the “Seminole” parameterization.
In the conventional parameterization of the Woods-Saxon potential, as well as in our work, the size
of nuclear potential is calculated as

R = RC = R0 A1/3 , RSO = R0,SO A1/3 , (3.56)

in terms of the parameters R0 and R0,SO , which are constant over the nuclear chart.
The surface diffuseness for both the central and spin-orbit potential is assumed to be constant and
size independent
a = aSO = const. (3.57)
While the isospin dependence of the Coulomb force is obvious, the behavior of the effective nuclear
potential on the isospin of the nucleon t and the core T0 has to be introduced phenomenologically. We
adopted the suggestion by Lane [6] to introduce an isospin dependence to the potential by the lowest
order isospin invariant term
 
4κ
V = V0 1 − ht · T0 i . (3.58)
A
We chose the “minus” sign so that the parameter κ will be consistent with conventions. For the
ground-state of a nucleus, the isospin quantum number is T = |Tz | = |N − Z|/2, which together with the
relation t + T0 = T leads to

 3 N =Z
−4ht · T0 i = ±(N − Z + 1) + 2 N > Z , (3.59)
±(N − Z − 1) + 2 N < Z


where here and below we use upper sign for the proton and the lower sign for the neutron. Traditionally,
the isospin-dependence of the Woods-Saxon potential had been parameterized by the expression.
 
(N − Z)
V = V0 1 ± κ (3.60)
A
For heavy nuclei with large neutron excess, the difference between the two definitions is small. However,
the definition of Eq. 3.59 leads to significantly different predictions in lighter nuclides around N=Z.
3 Nuclear Shell Model 46

Table 3.2: Commonly used Woods-Saxon parameter sets in the literature.

Parameterization V0 [M eV ] κ R0 [fm] a [fm] λ R0SO [fm]
’Rost’[7] n 49.6 0.86 1.347 0.7 31.5 1.28
p 1.275 17.8 0.932
’Optimized’[8] n 49.6 0.86 1.347 0.7 36 1.30
p 1.275 36 1.30
’Universal’[3] n 49.6 0.86 1.347 0.7 35 1.31
p 1.275 36 1.32
’Chepurnov’[9] 53.3 0.63 1.24 0.63 23.8 ∗ 1.24
’Wahlborn’[10] 51 0.67 1.27 0.67 32 1.27

Existing parameterizations
A number of parameterizations and parameter sets are available in the literature. We list the most
commonly used ones in Table 3.2. All of these parameterizations use Eq. (3.60) for the isospin dependence
of the nuclear potential. In contrast to the “Seminole” parameterization, the same isospin dependence
enters into spin-orbit term via Ṽ = λV .
The “Rost” parameters [7] were determined from the orbital energies of 208 Pb. The “Optimized” [8]
parameter set took the central potential parameters from “Rost”, but changed the spin-orbit interaction
in order to improve predictions for high-spin spectra in the lead region. A further refinement of these
parameters was introduced as the “Universal” parameter set, which improved the description of high spin
states in 146 Gd [3]. The “Chepurnov” parameterization introduced an additional isospin-dependence for
the spin-orbit interaction in deviation from the commonly used proportionality to the central potential.
3 Nuclear Shell Model 47

3.6 Homework problems

Exercise 1:
Show that J 2 , l2 , s2 and Jz , where J = l + s, commute with the Hamiltonian

p2
H= + V (r) + Vls (r)l · s
2µ
What does this implies?
Show that under a parity transformation a particle with angular momentum l and total spin j carries
a parity π = (−1)l .
Hint: [j (or jz ), H] = 0, [l (or lz ), H] 6= 0.

Exercise 2:
Draw the first 7 levels corresponding to the Hamiltonian

p2 1
H= + µωr2 + Vls l · s
2µ 2
where ~ω = 5 MeV and Vls = -0.5MeV.
Comment on the structure of the spectrum.

Exercise 3:
a) Determine the single-particle levels (including energy, l and j) for neutrons above the magic numbers
Z=50, N=82.
b) Do the same for protons.
Hint: You may just extract the excitation energies (relative the ground state) or the total energies of
the single-particle states from experimental data.

Exercise 4:
a) Determine the hole levels ( (including energy, l and j) ) in nuclei with protons below the magic
numbers Z=82, N=126.
b) Repeat for neutrons.
The nuclear experimental spectra are found in the site
http : //www.nndc.bnl.gov/nudat2/
 Bibliography

[1] M. Goeppert-Mayer, Phys. Rev. 75, 1969 (1949).

[2] R.D. Woods and D.S. Saxon, Phys. Rev. 95, 577 (1954).
[3] J. Dudek, Z. Szymanski, T. Werner, A. Faessler and C. Lima: Phys. Rev. C26 1712 (1982)

[4] Z. Lojewski, B. Nerlo-Pomorska, K. Pomorski, J. Dudek, Phys. Rev. C 51, 601 (1995)
[5] H. Koura and M. Yamada, Nucl. Phys. A671, 96 (2000).
[6] A. M. Lane, Phys. Rev. Lett. 8, 171 (1962); Nucl. Phys. 35, 676 (1962).

[7] E. Rost, Phys. Lett. B 26 184 (1968)

[8] J. Dudek, A. Majhofer, J. Skalski, T. Werner, S. Cwiok and W. Nazarewicz: J. Phys. G 5,10 1359
(1979)
[9] V.A. Chepurnov, Yad.Fiz 955 (1967)

[10] J. Blomqvist and S. Wahlborn, Ark. Fys. 16 (1960) 543

Chapter 4
Magnetic resonances in nuclei

Charge particles in a magnetic field. Time dependent magnetic fields. Time-dependent perturbation
treatment. Rabi formula. Magnetic Resonance Imaging (MRI). Dipole magnetic moments in nuclei.
Schmidt values.

4.1 Charge particles in a magnetic field

Assume a nucleon in the presence of a magnetic field carrying only its intrinsic angular momentum, i.
e. its 1/2-spin. This would happen is the nucleon is trapped within the region where the experiment
is performed. For instance, a proton in some molecules forming a crystal, or a proton in a molecule of
human tissue, which is largely composed of water with two hydrogen atoms (where the nucleus is the
proton itself) in each H2 O (water) molecule.
Assuming also that the magnetic field applied externally has the form

B = B0 k (4.1)

where B0 is constant and k is the unit vector in the z-direction, the Hamiltonian is

H = −µ · B = −µz B0 (4.2)

where the magnetic moment is defined by,

gq
µ= s (4.3)
2mc
In contrast to what we presented above, the magnetic moment is now defined with dimensions. This
is an unfortunate change of notation. We keep the notation used in each field. In Nuclear Physics the
Hamiltonian is as in Eq. (4.3), while in the applications of magnetic resonances to be analyzed here we
will use the Hamiltonian (4.2).
Since only the intrinsic spin of the particle is considered, the g-factor in Eq. (4.3) is as the gs factor
above, but for clarity of presentation we give them again here. In the cases of interest in the applications
the g-factors are 
 2.00 electron
g= 5.58 proton (4.4)
−3.82 neutron


As before, q is the charge of the particle (q = −e for electron) and s = (sx , sy , sz ) are the Pauli matrices
given by,      
~ 0 1 ~ 0 −i ~ 1 0
sx = ; sy = ; sz = (4.5)
2 1 0 2 i 0 2 0 −1
The Hamiltonian becomes,
 
gq ω0 ~ 1 0
H = −µ · B = − B0 s · k = ω0 sz = (4.6)
2mc 2 0 −1

where
gq
ω0 = − B0 (4.7)
2mc
and the eigenvalues are,
       
1 ω0 ~ 1 0 ω0 ~ 0
H = ; H =− (4.8)
0 2 0 1 2 1
4 Magnetic resonances in nuclei 50

There are two stationary (i.e. time independent) states with energies

ω0 ~
E± = ± (4.9)
2
If the particle is in the state +, it will not decay unless a perturbation disturbs it. When it decays a
photon with energy E+ − E− = ~ω0 will be emitted which can be measured with great precision, thus
allowing one to determine precisely quantities like the g-factor.

4.2 Time dependent magnetic fields

A convenient way to perturb the system is by applying a weak and time-dependent magnetic field in the
x-direction. Rabi chose for this purpose the form B1 cos ωt ix . The perturbation will then vary from −B1
to +B1 as the time increases. The hope is that at a certain value of ω the transition will take place.
Notice that B1 has to be very small in comparison to B0 in order not to destroy the spectrum determined
by B0 (i.e. the levels E± ). The problem is then to solve the Hamiltonian

gqB1
H = ω0 sz − cos ωtsx (4.10)
2mc
gqB1
with ω1 = − , one gets
2mc
     
ω0 ~ 1 0 ω1 ~ 0 1 ~ ω0 ω1 cos ωt
H= + cos ωt = (4.11)
2 0 −1 2 1 0 2 ω1 cos ωt −ω0

where |ω1 |  |ω0 |. One has to use the time-dependent Schrödinger equation, i.e.

dΨ(t)
HΨ(t) = i~ (4.12)
dt

Time-dependent perturbation treatment

Since B1 is very small the solution Ψ(t) should not be very different from the solution corresponding to
B1 = 0. We will therefore solve first the case B1 = 0, i. e.
    
~ ω0 0 a(t) ȧ(t)
= i~ (4.13)
2 0 −ω0 b(t) ḃ(t)

da(t)
where ȧ(t) = . One thus has
dt

~ da(t) 
ω0 a(t) = i~  a(t) = a(0) e−iω0 t/2


2 dt

=⇒ (4.14)
 ~
 − ω0 b(t) = db(t) 
b(t) = b(0) eiω0 t/2
i~

2 dt
The general case is
    
~ ω0 ω1 cos ωt a(t) ȧ(t)
= i~ (4.15)
2 ω1 cos ωt −ω0 b(t) ḃ(t)

since |ω1 |  |ω0 |, one proposes as solution

   −iω t/2 
a(t) e 0 c(t)
= (4.16)
b(t) eiω0 t/2 d(t)

which contains the main term explicitly.

4 Magnetic resonances in nuclei 51

The Schrödinger equation becomes

e−iω0 t/2 c(t)

  
~ ω0 ω1 cos ωt
2 ω1 cos ωt −ω0 eiω0 t/2 d(t)
−i ω20 e−iω0 t/2 c(t) + e−iω0 t/2 ċ(t)
 
= i~ ˙ (4.17)
i ω20 eiω0 t/2 d(t) + eiω0 t/2 d(t)

with cos ωt = eiωt/2 + e−iωt/2 /2,

+ ei(ω0 −ω)t d(t)

    i(ω +ω)t  
ċ(t) ω1 e 0
i ˙ =  −i(ω
0 −ω)t
(4.18)
+ e−i(ω0 +ω)t c(t)

d(t) 4 e

The idea is to change ω in the perturbation term B1 cos ωt such that ~ω0 ≈ ~ω. Since ω0 is large, the
highly oscillating functions e±i(ω0 +ω)t can be neglected. One thus gets
 ω1 i(ω0 −ω)t
 iċ(t) =
 e d(t)
4
(4.19)
˙ = ω1 e−i(ω0 −ω)t c(t)
 id(t)

4
which is a coupled set of two first order differential equations. To solve it one transforms it in a second
order differential equation as follows.
 ω1 i(ω0 −ω)t h ˙
i
 ic̈(t) =
 e i(ω0 − ω)d(t) + d(t)
4
(4.20)
ω h
¨ = 1 e−i(ω0 −ω)t − i(ω0 − ω)c(t) + ċ(t)
 id(t)

i
4
and replacing c(t) and ċ(t) from Eq. (4.19)

¨ ω1 −i(ω0 −ω)t h 4i i(ω0 −ω)t ˙ ω1 i(ω0 −ω)t 

id(t) = e − i(ω0 − ω) e d(t) + e d(t)
4 ω1 4i
 ω 2
˙ −i 1
= (ω0 − ω)d(t) d(t)
4
 2
¨ + i(ω0 − ω)d(t)
d(t) ˙ + ω1 d(t) = 0 (4.21)
4
which has the solution
1p
d(t) = Ae−i(ω0 −ω)t/2 sin Ωt, Ω= (ω0 − ω)2 + (ω1 /2)2 (4.22)
2
where A is a constant which is determined by the normalization condition, i.e.,
 

c(t) 2 2
c∗ (t), d∗ (t) = c(t) + d(t) = 1 (4.23)
d(t)

One proceeds in the same fashion with c(t) to obtain

s !
ω0 − ω1 i(ω0 −ω)t/2 ω12
c(t) = 2A e − sin Ωt − i 1 + cos Ωt (4.24)
ω1 4(ω0 − ω)2

Rabi formula

We have assumed that before the perturbation the system is in the state (+), i.e.

c(0) = 1; d(0) = 0 (4.25)

2 2 2
c(0) + d(0) = c(0) =1 (4.26)
4 Magnetic resonances in nuclei 52

Signal

E0

Figure 4.1: The resonant form of the signal as the energy E, corresponding to the weak magnetic field
B1 , approaches the energy E0 induced by B0 . The width of the resonance is Γ.

2
From c(0) = 1, and after some algebra, one gets,

2 (ω1 /2)2
A = (4.27)
(ω0 − ω)2 + (ω1 /2)2

and the probability that the transition takes place, i.e. that the system is in the state (−) is

2 (ω1 /2)2
d(t) = sin2 Ωt (4.28)
(ω0 − ω)2 + (ω1 /2)2

and a resonance occurs when ω = ω0 . Eq. (4.28) is the Rabi’s formula.

Nuclear magnetic resonance was first described and measured in molecular beams by Isidor Isaac Rabi
in 1938. In 1944, Rabi was awarded the Nobel Prize in physics for this work.
http://www.nobelprize.org/nobel_prizes/physics/laureates/1944/

4.3 Nuclear magnetic resonance (NMR)

Nuclear magnetic resonance (NMR) is a physical phenomenon in which magnetic nuclei in a magnetic
field absorb and re-emit electromagnetic radiation. This energy is at a specific resonance frequency which
depends on the strength of the magnetic field and the magnetic properties of the isotope of the atoms.
One sees that A shows a form similar to that of the Breit-Wigner formula

(ω1 /2)2 ~2 (Γ1 /2)2

= (4.29)
(ω0 − ω)2 + (ω1 /2)2 ~2 (E0 − E)2 + (Γ1 /2)2
from which one can define E = ~ω0 as the magnetic resonance energy and Γ1 = ~ω1 as the the width.
In Fig. 4.1 the form of the signal resulting from this expression is shown. The signal-energy plot shown
in Fig. 4.1 has been used to investigate the inner structure of materials. Many scientific techniques
exploit NMR phenomena to study molecular physics, crystals, and non-crystalline materials through
NMR spectroscopy. NMR allows the observation of specific quantum mechanical magnetic properties
4 Magnetic resonances in nuclei 53

Table 4.1: the gyromagnetic ratio γ of typical nuclei.

Nucleus Spin γ(MHz/T) Natural Abundance

1
H 1/2 42.576 99.985
2
H 1 6.536 0.015
13
C 1/2 10.705 1.11
14
N 1 3.077 99.63

of the atomic nucleus. In particular, it is used in Medicine to image nuclei of atoms inside the body
(magnetic resonance imaging (MRI)).
In all applications of MRI one uses SI units and introduces the Bohr magneton
q~
µB = (4.30)
2mc
In these units the frequency becomes
gqB0 gµB B0
ω0 = = = γB0 (4.31)
2mc ~
This is known as the Larmor Equation. As already mentioned, for electrons it is g = 2.00 and µB =
5.79×10−5 eV/T, where the unit Tesla is 1T = 104 gauss. Remember ~c ≈ 200MeVfm. The precise value of
~ is ~ = 6.58×10−22 MeVsec. For protons g = 5.58 (as also already mentioned) and µN = 3.15×10−8 eV/T.
In practical applications, the frequency is similar to VHF and UHF television broadcasts (60−1000 MHz).

Magnetic Resonance Imaging (MRI)

An MRI machine uses a powerful magnetic field to align the magnetization of some atoms in the body,
and radio frequency fields to systematically alter the alignment of this magnetization. This causes the
nuclei to produce a rotating magnetic field detectable by the scanner and this information is recorded
to construct an image of the scanned area of the body. Strong magnetic field gradients cause nuclei at
different locations to rotate at different speeds. 3-D spatial information can be obtained by providing
gradients in each direction.
MRI provides good contrast between the different soft tissues of the body, which make it especially
useful in imaging the brain, muscles, the heart, and cancers compared with other medical imaging tech-
niques such as computed tomography (CT) or X-rays. Unlike CT scans or traditional X-rays, MRI uses
no ionizing radiation.
This last is a very important point, since it implies that no damage of the human tissue is associated
with MRI.
All isotopes that contain an odd number of protons and/or of neutrons have an intrinsic magnetic
moment. The most commonly studied nuclei are 1 H and 13 C (both with spin 1/2). Hydrogen is highly
abundant, especially in biological systems in water (H2 O) and lipid (CH2 ) molecules. It is the nucleus
most sensitive to NMR signal (apart from 3 H which is not commonly used due to its instability and
radioactivity). Proton NMR produces narrow chemical shift with sharp signals. Fast acquisition of
quantitative results (peak integrals in stoichiometric ratio) is possible due to short relaxation time. The
1
H signal has been the sole diagnostic nucleus used for clinical MRI. Nuclei from isotopes of many other
elements (e.g., 2 H, 6 Li, 10 B, 11 B, 14 N, 15 N, 17 O, 19 F, 23 Na, 29 Si, 31 P, 35 Cl, 113 Cd, 129 Xe, 195 Pt) have
been studied by high-field NMR spectroscopy as well. Most even-even nuclei have zero nuclear magnetic
moments, and they also have zero magnetic dipole and quadrupole moments. Hence, such nuclides do
not exhibit any NMR absorption spectra.
The biological abundance describes the fraction of certain atom in the human body. The biological
abundances of H, O, C and N in the body are 0.63, 0.26, 0.094, 0.015, respectively.
http://www.cis.rit.edu/htbooks/mri/

Electron spin resonance

The magnetic moment of an electron amounts to 658 times the magnetic moment of a proton. The
electron does not contribute to the signal in MRI, though it can undergo a similar process as the nucleus,
4 Magnetic resonances in nuclei 54

called electron spin resonance (ESR). ESR is a related technique in which transitions between electronic
spin levels are detected rather than nuclear ones. The basic principles are similar but the instrumentation,
data analysis, and detailed theory are significantly different.
The electron gyromagnetic ratio is given by 28024.95266(62) MHz/T. The magnetic field produced by
an electron is much stronger than that produced by a proton. However, in most substances the electrons
are paired, resulting in a weak net magnetic field. There is a much smaller number of molecules and
materials with unpaired electron spins that exhibit ESR (or electron paramagnetic resonance (EPR))
absorption than those that have NMR absorption spectra.

Nuclear magnetic resonance quantum computer

NMR quantum computing uses the spin states of molecules as qubits. Operations are performed on the
ensemble through radio frequency pulses applied perpendicular to a strong, static field, created by a
very large magnet. Some early success was obtained in performing quantum algorithms in NMR systems
due to the relative maturity of NMR technology. For instance, in 2001 researchers at IBM reported the
successful implementation of Shor’s algorithm in a 7-qubit NMR quantum computer
(http://www.nature.com/nature/journal/v414/n6866/full/414883a.html).

4.4 Magnetic fields and magnetic moments

A nucleon moving in a single-particle state outside a central potential, as discussed above, will be affected
by the presence of an external magnetic field. The corresponding Hamiltonian is,
q
H=− µ·B (4.32)
2mc
where q is the effective charge of the nucleon, B is the magnetic field and µ is the dimensionless nuclear
dipole moment defined as,
µ = gl ˆl + gs ŝ (4.33)
The effective charge should be q = 1.0 e (e is the absolute value of the electron charge) for protons and
0 for neutrons. However, its value is taken to be about 1.5 e for protons and about 1.0 e for neutrons
(these are illustrative values that can vary in different nuclear regions, i. e. for different values of N and
Z). The reason why the effective charge was introduced is that the odd nucleon affects the core and it
has been shown that its influence can be taken into account by the effective charge.
The magnetic moments can be measured with great precision, thus providing precise value for the
g-factors also. These are given by,
 
1 proton 5.58 proton
gl = gs = (4.34)
0 neutron −3.82 neutron

Dipole magnetic moments in nuclei

When the magnetic field is applied the energies observed experimentally are quantized according to the
allowed angular momenta in Eq. (4.33). To measure the dipole magnetic moment µ one chooses the
maximum splitting of the levels. One sees from Eq. (4.32) that the maximum effect of the magnetic field
would be induced by the maximum alignment of µ and B. This is what one chooses experimentally.
Classically this occurs when jz = j (since B is in the z-direction). In Quantum Mechanics one has to
choose the projection m of the total angular momentum such that m = j. Therefore one defines the
dipole magnetic moment as
µ = hjm = j|gl ˆlz + gs ŝz |jji (4.35)
notice that µ is just a number without dimensions.
Since l = j − s one can write
µ = hjj|gl ĵz + (gs − gl )ŝz |jji (4.36)
To calculate the values obtained by the application of the operator ŝz upon the state |jji, we expand
this state in terms of the eigenvectors of ŝz , i. e.
X
|jmi = hlml 1/2ms |jmi|lml 1/2ms i (4.37)
ml ms
4 Magnetic resonances in nuclei 55

and one obtains

X  
µ = hjj| hlml 1/2ms |jji gl j + (gs − gl )ms |lml 1/2ms i
ml ms
X
hlml 1/2ms |jji2 gl j + (gs − gl )ms
 
=
ml ms

It is ml = j − ms and ms = ±1/2, ml = j ∓ 1/2. Therefore

 
D 1 1 1 E2 gs − gl
µ = l, j − , , jj gl j + (4.38)
2 2 2 2
 
D 1 1 1 E2 gs − gl
+ l, j + , , − jj gl j −
2 2 2 2

Schmidt values of single-particle states

The general expression for the magnetic moments corresponding to a nucleon moving outside a central
field is given by Eq. (4.38). There are two different possibilities, namely a) l = j + 1/2; b) l = j − 1/2.

a) j = l − 1/2
D 1 1 1 1 E2 1
j + , j − , , jj =
2 2 2 2 2(j + 1)
D 1 1 1 1 E2 2j + 1
j + , j + , , − jj =
2 2 2 2 2(j + 1)

j
µ = gl j − (gs − gl ) (4.39)
2(j + 1)

b) j = l + 1/2
D 1 1 1 1 E2
j − , j − , , jj = 1
2 2 2 2
D 1 1 1 1 E2
j − , j + , , − jj = 0
2 2 2 2
µ = gl j + (gs − gl )/2 (4.40)

Our assumption that the nucleon moves around the core without farther disturbances (for instance
without exciting other nucleons in the core) is equivalent to assume that the nucleon moves in a single-
particle state. If this is valid, then one can estimate probable values of the angular momenta l and j (we
will perform this task in this Course). Since also the Schmidt values for the magnetic moment should
be valid, then one can compare these magnetic moments with experiment to probe the models used to
infere the angular momenta as well as the single-particle assumption that this implies. We will attest the
success of this in the exercises.
4 Magnetic resonances in nuclei 56

4.5 Homework problems

Exercise 1:
a) Calculate the spectrum corresponding to a spin 1/2 particle with magnetic moment µ ~ = gq/(2mc)~s
~ = Bx~i + By~j where Bx and By are constants.
situated in a magnetic field B
b) Which is the probability that in the lowest state thus calculated the spin projection of the particle
is measured to be sz =-1/2.

Exercise 2:
a) In the Rabi formula one has to assume |B1 /B0 |  1. Why?.
b) Show that d(t) (Eq. (4.22) in Chapter 3) is solution of Eq. (4.21).

Exercise 3 (not mandatory):

Consider medical applications of Magnetic Resonance Imaging (MRI). In these applications,
a) Only protons are considered, while other heavier isotopes, like 16 O, are neglected. Why?.
b) Why are the interferences of electrons also neglected?.
c) The intrinsic spin of the proton is considered to be j = s. Why?.

Exercise 4:
Which is the width of a MRI resonance if the variable field perpendicular to B0 =1 T is B1 =10 gauss?.
How much changes this width if B0 =2 T?.

Exercise 5:
Evaluate ω0 for a proton in a magnetic field B0 =1.0 T (which is the most common value used
in medical applications of MRI). Compare this frequency with the one corresponding to the x-rays in
computed tomography (CT), with an energy of 10 keV, and the γ-rays of radiation therapy, with an
energy of 1 MeV.
Exercise 6:
Evaluate the magnetic dipole moment (Schmidt value) corresponding to
a) A proton moving in an orbital l = 2, j = 5/2.
b) A neutron moving in an orbital l = 1, j = 1/2.

Exercise 7:
a) The magnetic dipole moments corresponding to the ground states of the nuclei 17 F8 and 41 Sc20
are observed to be 4.72 and 5.53, respectively. Which are the corresponding values of l and j?.
b) Which values of l and j would have the ground states of 17 O9 and 41 Ca21 , for which the magnetic
dipole moments are measured to be -1.83 and -1.59, respectively?.
Chapter 5
Rotational model

Rotational model

5.1 Rotational model

So far we have considered the nucleus as a Fermi gas consisting of neutrons and protons moving freely
under the influence of the shell model central potential. We have indicated that the success of the shell
model was in the beginning considered a wonder since the very short range of the nuclear force acting
upon the tightly packed nucleons make it difficult to understand their free motion. There are arguments
to justify this, but still one would rather think that those characteristics are more suitable to generate
a collective behavior where all nucleons move together as a whole preserving the nuclear volume. These
considerations gave rise to the so-called liquid drop model in which the individual nucleons do not play
any role. It is the nucleus as a whole that determines its dynamics. Like a liquid drop, the surface of the
nucleus vibrates and rotates generating bands of quantized states. An even more limited collective model
considers the nucleus as a rigid body with fixed center of mass. The only possible motions of such an
object are rotations. This is the rotational model, where the conglomerate of individual nucleons form a
compact entity, like a top or a rugby ball.
From classical mechanics it is known that three angles are needed to define the position of a rigid
body with fixed center of mass. These are called Euler angles. One chooses an intrinsic axis system,
called (x’,y’,z’) in Fig. 5.1, and determines the direction of one of the axis, say z’, with respect to the
laboratory system (x,y,z). This operation requires two of the Euler angles. The third one is used to
obtain the orientation of the body along the intrinsic axis z’. The Euler angles are denoted by (θ, φ, ϕ).
The energy of the rotating rigid body, with the center of mass fixed at the center of coordinates, is

Jx20 Jy20 J 20
E= + + z (5.1)
2Ix0 2Iy0 2Iz0

where Ix0 is the x0 component of the moment of inertia and Jx0 is the corresponding angular momentum
component.
We will assume that the rigid body has cylindrical symmetry along the z’ axis. Therefore the com-
ponent Jz0 of the angular momentum, which is usually denoted by the letter K, is conserved. This
symmetry also implies that Ix0 = Iy0 . We will use the symbol I to denote this moment of inertia.
In the intrinsic system the rigid body is at rest and here the symmetry is only cylindrical. But in the
laboratory system the symmetry is spherical because the rigid body rotates in all directions and there
is no way to distinguish one of these directions from another. Therefore the angular momentum J is
conserved in the laboratory system and that implies, as we have seen in Chapter 1, that J 2 and Jz are
good quantum numbers which can be used to label the rotational state. The quantum number associated
with Jz is usually called M . The state thus carries J, M, K as quantum numbers, i. e. it can be written
as |JM Ki or,
J
DM K (θ, φ, ϕ) = hθφϕ|JM Ki (5.2)
which are called ”d-functions”. They satisfy the eigenvalue equations,

J 2 DM
J 2 J
K (θ, φ, ϕ) = ~ J(J + 1)DM K (θ, φ, ϕ),
J J
Jz DM K (θ, φ, ϕ) = ~M DM K (θ, φ, ϕ), (5.3)
J J
Jz0 DM K (θ, φ, ϕ) = ~KDM K (θ, φ, ϕ)

In classical mechanics one can distinguish whether an spherically symmetric body rotates or not, but
in quantum mechanics all directions are the same and the body appears to be at rest. Therefore there is
5 Rotational model 58

M J axis
etry
sym
m z'

Figure 5.1: A cylindrically symmetric rotator. The symmetry axis is z 0 . The projection of the angular
momentum upon this axis is K, and upon the z-axis in the laboratory frame is M .

no rotational energy associated to such a system or, in general, to degrees of freedom corresponding to a
cylindrical symmetry. The Hamiltonian (5.1) thus becomes,
Jx20 + Jy20 J 2 − Jz20
H= = (5.4)
2I 2I
The eigenfunctions of this Hamiltonian satisfying all the symmetries of the system, including the cylindri-
cal symmetry, the reflection symmetry with respect to the (x’,y’) plane and the parity symmetry (notice
that the Hamiltonian is invariant under the parity operation of changing r by −r) is,
J J J


hθφϕ|JM Ki = c DM K + (−1) DM −K (5.5)

where c is a normalization constant. The corresponding eigenenergies are, from Eq. (5.4),
J(J + 1) − K 2
E(J, K) = ~2 (5.6)
2I
Since J ≥ K and positive for a given K ≥ 0 there is a band of states with energies proportional to
J(J + 1). This is call ”rotational band”. The lowest lying of these bands is the one corresponding to
K = 0, and its lowest state, i. e. E(0, 0) = 0, is the ground state of the rotational nucleus. Therefore this
band is called ground state band. From Eq. (5.5) one sees that for the ground state band hθφϕ|JM Ki
vanishes if J is odd. Therefore the members of the ground state band have even J and parity (−1)J =
+1 (since parity is conserved) and the energy is
J(J + 1)
E(J, 0) = ~2 (5.7)
2I
This is a very characteristic spectrum and many nuclei follows it. An example is the nucleus 238 P u which
spectrum is shown in Fig. 5.2.
An important sign that a spectrum corresponds to a ground state band is that the relation between
the energies of the first 4+ and 2+ states should be 10/3, as shown by Eq. (5.7). In general, the
levels of a rotational ground band are related to the energy of the first excited state 2+ by the relation
E(J, 0) = E(2, 0) J(J + 1)/6. The agreement between theory and experiment can be excellent, as is seen
in Fig. 5.2. The moment of inertia of a rotational nucleus can be extracted from the experimental energy
E(2, 0), since
J(J + 1)
I = ~2 . (5.8)
2E(2, 0)
5 Rotational model 59

+
5 1 4 8

+
3 0 3 6

+
1 4 6 4

+
4 4 2
+
0 0
2 3 8
P u 9 4

Figure 5.2: Ground rotational band (i. e. K=0) of 238 P u. Energies are in keV. The rotational energies
E(J, 0) = E(2, 0) J(J + 1)/6, where E(2, 0) = 44keV in this case, are E(4, 0)=147 keV, E(6, 0)=308 keV
and E(8, 0)=528 keV.

For the K > 0 bands one can show that K must be even and J=K, K+1, K+2,... The lowest of the
states of a band is called bandhead, and it corresponds to J=K. Therefore the bandhead energy Ebh is
given by,
K
Ebh (K) = ~2 (5.9)
2I
The rotational model has been very successful to explain the spectra of a large number of nuclei. It
was formulated by A. Bohr and B. Mottelson in the beginning of the 1950’s and it was corroborated
experimentally by J. Rainwater. They were awarded the 1975 Nobel Prize in Physics for this work.
This model complements the shell model since it works very well in the middle of the major shells of
Fig. 3.2, i. e. with values of the number of neutrons and protons which are far from magic numbers.
However the shell model is, more than a model, a procedure to obtain a very good representation to solve
the nuclear many-body problem. Therefore there have been many attempts to explain rotational spectra
by using the shell model. This, which is even been pursued at present, has had a great importance in the
understanding of nuclear correlations.
One can separate out the deformed nuclei from the others by making use of the ratio excitation
energies
E4
R42 = . (5.10)
E2
It is a good indicator of the character of the nucleus and has the value R42 = 10/3 for an axial rotor.
A histogram of R42 for all the nuclei for which the energies are known is shown in Fig. 5.3. There is a
sharp peak around the rotor value.

5.2 Deformed shell model (Nilsson model)

The spherical shell model can explain many features of spherical nu- clei, but needs modifying to describe
nuclei with many nucleons outside a closed shell. The residual interaction between these many valence
nucleons may be more simply described by a deformed potential. For nuclear rotation to be observable,
the nuclei have to be non-spherical, so that they have a preferred axis.
Quadrupole (λ = 2) deformations can give rise to asymmetric shapes. These triaxial distortions are
governed by the γ shape degree of freedom, and this describes a stretching/squashing effect at right angles
5 Rotational model 60

100

80
Number of nuclei

60

40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4
R42

Figure 5.3: Distribution of nuclei with respect to deformation indicator R42 . Taken from GF Bertsch,
arXiv:1203.5529.

Figure 5.4: The different nuclear shapes that can be parametrised by spherical harmonic functions, where
λ characterises the different orders of the corresponding distributions.
5 Rotational model 61

to the major nuclear axis. Gamma is measured in degrees, where γ = 0◦ and γ = 60◦ correspond to
prolate and oblate shapes respectively. Completely triaxial shapes have γ = 30◦ .
The model that describes axially symmetric nuclei is called the Deformed Shell Model. In this model
the Schrödinger equation is solved using the potential that describes the shape of the nucleus. An result
of the deformation is that the orbital angular momentum, l, and the intrinsic spin, s, are no longer good
quantum numbers and thus, states with different l-values but the same parity can mix. The energy of the
states now depends on the component of the single-particle angular momentum (j) along the symmetry
axis, which is denoted by Ω. For each orbital with angular momentum j, there are 2j+1 values of Ω
(= mj in the absence of other couplings). However, levels with +Ω and Ω have the same energy due to
the reflection symmetry of axially symmetric nuclei, so that each state is now doubly degenerate, i.e. two
particles can be placed in each state. For example the f7/2 orbital can have |Ω| equal to 7/2, 5/2, 3/2
and 1/2. The ordering of these Ω levels depends on the particular shape of the nucleus since the lowest
in energy is the orbital which interacts (overlaps) the most with the nuclear core. For prolate shaped
nuclei the states with the lowest Ω values are the most tightly bound, whereas for oblate shaped nuclei,
the states with the highest Ω occur lowest in energy. Such deformed shell model calculations were first
performed in 1955 by Nilsson with an anisotropic harmonic oscillator potential and the calculated states
(called Nilsson orbitals) are labelled by Ω[N nz Λ] where N is the total oscillator shell quantum number
and determines the parity, given by (−1)N . Lambda is the projection of the particle orbital angular
momentum, l, on the nuclear symmetry axis, and nz is the number of oscillator shell quanta along the
direction of the symmetry axis.
5 Rotational model 62

5.3 Homework problems

Exercise 1:
Determine the moment of inertia from the experimental spectrum corresponding to the rotational
ground state band in
(a) 238 Pu.
(b) 168 Yb.
Hint: Experimental data could be found in
http://www.nndc.bnl.gov/nudat2/
Chapter 6
Two-particle states

Two-particle states. Parity of n-particle states. Isospin. Sum of isospins. Consequences of antisymme-
try.

In order to evaluate the two-particle matrix elements we have to extend the formalism to r-space. To
this end we will assume that the particles move in the shells |pi = |np lp jp mp i and |qi = |nq lq jq mq i of
the chosen single-particle representation. Since j = l + s the corresponding single-particle wave function
is (Chapter 1),  
hr|pi = Rnp lp jp (r) Ylp (r̂)χ1/2 j m (6.1)
p p

The two-particle representation consists of the product of the one-particle states, i. e.

h12|pqi = hr1 |pihr2 |qi (6.2)

and the corresponding antisymmetric wave function is.

h12|pqia = N (h12|pqi − h21|pqi) (6.3)

√
where N is a normalization constant, which in this case is N = 1/ 2. Writing explicitly the brackets in
terms of the radial and spin coordinates one gets,

1 h    
Ψa (pq; r1 r2 ) = √ Rp (r1 ) Ylp (r̂1 )χ1/2 p Rq (r2 ) Ylq (r̂2 )χ1/2 q
2
    i
− Rp (r2 ) Ylp (r̂2 )χ1/2 p Rq (r1 ) Ylq (r̂1 )χ1/2 q (6.4)

In this equation each index carries all the quantum numbers, e. g. {p ≡ np lp jp mp } although in the
angular momentum function Ylp the orbital angular momentum is shown explicitly. The quantum numbers
associated to Eq. (6.4) are {np lp jp mp nq lq jq mq }. This is called the m-scheme. In this scheme the Pauli
principle forbids all quantum numbers to be the same since if q = p, then Ψa (pq; r1 r2 ) = 0. .
We have also seen that one can choose another coupling scheme, in which the two particles carry total
angular momentum J = jp + jq with projection M = mp + mq , such that |jp − jq | 6 J 6 jp + jq and
−J 6 M 6 J. This is called coupled-scheme. In this scheme the two-particle wave function reads,
h h    i
Ψa (pq, JM ; r1 r2 ) = N Rp (r1 )Rq (r2 ) Ylp (r̂1 )χ1/2 j Ylq (r̂2 )χ1/2 j
p q JM
h    i i
− Rp (r2 )Rq (r1 ) Ylp (r̂2 )χ1/2 j Ylq (r̂1 )χ1/2 j (6.5)
p q JM

The quantum numbers associated to this wave function are {np nq lp lq jp jq JM }. There are eight quantum
numbers in both schemes, as it should be since the quantum numbers express symmetries of the system
that do not depend upon the chosen scheme.
In the coupled-scheme the quantum numbers mp and mq do not appear (remember that the radial
wave functions Rp (r) do not depend on mp ). Yet in the shells jp and jq one can place 2jp + 1 and
2jq + 1 particles, respectively. Therefore in the coupled-scheme all quantum numbers can be the same.
This can be seen from Eq. (6.5) since it is not apparent that the wave function vanishes when p = q.
To see how the Pauli principle works in this case it is convenient to exchange the angular momentum
coupling of the second term in Eq. (6.5), from (jp jq ; JM ) to (jq jp ; JM ). One gets (see Exercise 4a of
Homeworkproblems 1),
h h    i
Ψa (pq, JM ; r1 r2 ) = N Rp (r1 )Rq (r2 ) Ylp (r̂1 )χ1/2 j Ylq (r̂2 )χ1/2 j
p q JM
h  i i
jp +jq −J
 
− Rp (r2 )Rq (r1 )(−1) Ylq (r̂1 )χ1/2 j Ylp (r̂2 )χ1/2 j (6.6)
q p JM
6 Two-particle states 64

For p = q it is,
h  i
(1 − (−1)2jp −J )
 
Ψa (pp, JM ; r1 r2 ) = N Rp (r1 )Rp (r2 ) Ylp (r̂1 )χ1/2 j Ylp (r̂2 )χ1/2 j (6.7)
p p JM

Since jp is a semi integer one gets (−1)2jp = -1. Therefore the Pauli principle requires that if p = q
the total angular momentum J should be even. One sees in the equation above that in this case the
normalization constant is N =1/2, and the antisymmetrized wave function becomes,
h    i
Ψa (pp, JM ; r1 r2 ) = Rp (r1 )Rp (r2 ) Ylp (r̂1 )χ1/2 j Ylp (r̂2 )χ1/2 j (6.8)
p p JM

Parity of n-particle states

We have seen in the previous Chapter that within the shell model the correlated two-particle states (that
is the eigenvectors of the total Hamiltonian) are expanded in terms of the basis states (6.4) in m-scheme
or (6.5) in coupled-scheme. In both cases the angular part of the wavefunction is a product  of the single-

particle Harmonic functions, e. g. in Eq. (6.4) this product appears as Ylp (r̂1 )χ1/2 j m Ylq (r̂2 )χ1/2 j m .
p p q q
The parity operation consists of making the transformation (r1 , r2 ) → (−r1 , −r2 ) and, as we have seen
in Chapter 1, this transformation
 adds a phase
 (−1)l in the Harmonic function Ylm . Therefore the term
lp +lq
above becomes (−1) Ylp (r̂1 )χ1/2 j m Ylq (r̂2 )χ1/2 j m and the parity of the state is (−1)lp +lq . For
p p q q
instance, if a particle is in a state with jp = s1/2 and the other in a state jq = p3/2 the parity of the
possible states is −1. The possible total angular momenta are |1/2 − 3/2| ≤ J ≤ 1/2 + 3/2, that is the
possible states are 1− and 2− . In the same fashion, if p = q and lp = lq = 3, jp = jq = 5/2 the possible
states are 0+ , 2+ and 4+ .
In the n-nucleon case the basis consists of the product of n single-particle states (since
 they are eigen-

vectors of the free Hamiltonian H0 ), The Harmonic functions in this product is given by Ylp1 (r̂1 )χ1/2 j m
p1 p1
   
Ylp2 (r̂2 )χ1/2 j m ... Ylpn (r̂n )χ1/2 j m and, as before, the parity is (−1)l1 +l2 +...+ln .
p2 p2 pn pn

6.1 Isospin

Except the electric charge, protons and neutrons are very similar to each other. Not only they possess
almost the same mass (Mn /Mp = 1.0014), but they also show a far reaching symmetry with regard to
the nuclear interaction. Thus, experimental results extracted from proton-proton and proton-neutron
scattering show that the nuclear forces are virtually the same for the proton-proton and neutron-proton
systems. The properties of mirror nuclei, i. e. nuclei in which the number of protons and neutrons are
exchanged retaining the same nuclear number A, are also very similar to each other. In particular the
energy spectra of these nuclei are very much alike if effects induced by the Coulomb force are subtracted,
as we will see.
One can consider neutrons and protons to be two states of the same particle, the nucleon. The nucleon
is determined by the strong nuclear force. The Coulomb interaction is much weaker and has a very long
range. At difference with the nuclear case, the Coulomb interaction is very well known and it is relatively
easy to include its influence independently of the nuclear force.
The two states of the nucleon suggest a similarity with the spin and its two projections. This similarity
prompted Heisenberg in 1932 (soon after the neutron was discovered) to introduce a new quantum number,
the isospin t̂, with value 1/2 and two projections in an abstract space, the isospin space. The projection
up of the isospin, |νi, is a neutron and the isospin down, |πi, a proton, i. e. for neutrons t̂z |νi = 21 |νi
and for protons t̂z |πi = − 12 |πi. The reason why one has taken the convention of given the value +1/2
to the neutron eigenvalue of t̂z and −1/2 to proton, is that for the vast majority of nuclei it is N ≥ Z.
This implies that the total isospin projection of the nucleus (N, Z) is (N − Z)/2 ≥ 0.
One defines the isospin projection in terms of the Pauli matrices following the spin relations, i. e.
     
1 1 1 0 1 0
t̂z = τ̂z = ; |νi = , |πi = (6.9)
2 2 0 −1 0 1
6 Two-particle states 65

and the isospin is defined also following the spin algebra, i. e. t̂2 = t̂2x + t̂2y + t̂2z . The components
(tx , ty , tz ) satisfy the same conmutation relations as the spin and, therefore, one can choose the eigenvalues
(t, tz ) = (1/2, tz ) to label the isospin states.
For the operator t̂2 it is,
3
t̂2 |1/2 tz i = t(t + 1)|1/2 tz i = |1/2 tz i (6.10)
4
and for tˆz ,
tˆz |1/2 tz i = tz |1/2 tz i (6.11)
One can define a charge operator in terms of the z-component of the isospin as,
     
1 1 1 0 1 0 0 0
Q̂ = (1 − τ̂z ) = − = (6.12)
2 2 0 1 0 −1 0 1

One finds that Q̂ is indeed the charge operator since its eigenvalues are the same as those of t̂z , that is
       
1 1 0 0
Q̂|νi = Q̂ = eν ; Q̂|πi = Q̂ = eπ (6.13)
0 0 1 1
and the eigenvalues are the charges, with the values eν =0 and eπ =1, as it should be.
Since the Hamiltonian conserves charge one has
[H, Q̂] = 0 =⇒ [H, t̂z ] = 0 (6.14)
i.e. tz is a constant of the motion and, therefore, a quantum number.
The single-particle wave function can then be labeled by the spatial-spin quantum numbers |nlsjmi
(where s = 1/2) and the isospin quantum numbers |ttz i (t = 1/2). The corresponding wave function is,
hr|nlsjmttz i = Rnlj (r) [Yl (r̂)χs ]jm τttz (6.15)
where τttz is the isospin state. It should be stressed that the isospin bears no relation to ordinary space.
Before finishing this Section it has to be emphasized that since the masses of the neutron and proton
are not exactly the same, isospin is not an exact symmetry, although for all practical purposes one
can consider isospin a good quantum number. The reason of this broken of symmetry is explained by
the theory behind strong interactions, Quantum Chromo Dynamics (QCD). This theory describes the
nucleon as been composed of three quarks, called up and down quarks, with different masses which are
not possible to measure directly because the theory predicts that free quarks cannot be observed. The
neutron contains 1 quark up and 2 down, while the proton contains 2 up and 1 down.

Ladder operators
In analogy with angular momentum one can define raising and lowering operators as
 
1 0 1
t+ = (τx + iτy ) = , (6.16)
2 0 0
and  
1 0 0
t− = (τx − iτy ) = . (6.17)
2 1 0
The commutation relations between the ladder operators and tz can easily be obtained as
[tz , t± ] = ± t± ,
(6.18)
[t+ , tz ] =2tz .
Direct application of the explicit matrix representations to the isospinors yields the relations (i.e., beta
decay process)
t+ |pi =|ni,
t+ |ni =0,
(6.19)
t− |ni =|pi,
t− |pi =0.
or
p
t± |ttz i = (t ∓ tz )(t ± tz + 1)|ttz ± 1i. (6.20)
6 Two-particle states 66

Sum of isospins

Since the isospin operators obey the angular momentum algebra (that is, the same commutation relations
as the angular momentum operators) all the studies done for the case of sum of angular momenta is valid
also here. Thus, given the isospins t1 and t2 , the two-nucleon state can be labeled by using the m-scheme,
i. e. (t1 t1z t2 t2z ), with t1 = t2 = 1/2. One can also use the coupled-scheme by introducing the total
isosspin T = t1 + t2 with z-projection Tz = t1z + t2z and the state can be labelled by (t1 t2 T Tz ). Both
schemes can be used to describe any state in the space of isospin. In particular one can write a state in
the m-scheme in in term of the states in coupled-scheme, as done in Eq. (1.91) of Chapter 1, by
X
|t1 t1z t2 t2z i = ht1 t1z t2 t2z |T Tz i|t1 t2 T Tz i (6.21)
T Tz

or the other way around, as done in Eq. (1.92) of Chapter 1,

X
|t1 t2 T Tz i = ht1 t1z t2 t2z |T Tz i|t1 t1z t2 t2z i (6.22)
t1z t2z

where ht1 t1z t2 t2z |T Tz i is the Clebsh-Gordan coefficient.

The triangular relation in this isospin case is, as usual, |t1 − t2 | 6 T 6 t1 + t2 and −T 6 Tz 6 T .
Since t1 = t2 = 1/2 one has T = 0, 1. The T = 0 (T = 1) excitation is called ”isoscalar” (”isovector”).
In general, with A nucleons with isospin ti (i = 1, 2...A), one has
A
X A
X
T= ti =⇒ Tz = tiz (6.23)
i=1 i=1

and with A = N + Z one gets

N −Z
T̂z |Ai = |Ai (6.24)
2
i.e. the isospin projection is (N − Z)/2.
We define
X
T± = ti± . (6.25)
i

It can be shown that one has

p
T± |T Tz i = (T ∓ Tz )(T ± Tz + 1)|T Tz ± 1i. (6.26)

Since Tz is bounded by the value of −T ≤ Tz ≤ T we have

T± |T ± T i = 0. (6.27)

Consequences of antisymmetry
The isospin symmetry that we are assuming implies that an A-body state defined by the space-spin
quantum numbers bears also the isospin quantum numbers and therefore the wave function of such an
state carries the quantum numbers

{(n1 l1 s1 j1 m1 t1 tz1 , n2 l2 s2 j2 m2 t1 tz1 , ..., nA lA sA jA mA tA tzA )}. (6.28)

The meaning of the isospin symmetry is that an A-nucleon state with isospin T will be present in all
nuclei with A nucleons satisfying A = N + Z. But this requires that all the other quantum numbers, i. e.
those corresponding to the space-spin degrees of freedom, are the same. Therefore generally the isospin
is a useful quantity only when the nucleons move all in the same orbits. This means that the core should
have N ≈ Z, but preferably it should be self conjugate, i. e. N = Z.
The antisymmetric wave function of two nucleons moving outside a frozen core is like in Eq. (6.6) but
including also the isospin component of the wave function. If the two nucleons move in the same shell,
as it was the case in Eq. (6.7), the wave function including isospin satisfies,

h12|(nlj)2JM (1/2 1/2)T Tz ia = N 1 − (−1)J+T h12|(nlj)2JM (1/2 1/2)T Tz i

(6.29)
6 Two-particle states 67

which shows that the wave function vanishes unless J + T is odd.

Since the space-spin part of the wave function is decoupled from the isospin part one can write,

h12|(nlj)2JM (t1 , t2 )T Tz ia = h12|(nlj)2JM ih12|(1/2 1/2)T Tz i a

(6.30)

That is, either the spatial-spin part is symmetric and the isospin part antisymmetric, or the other way
around. This is determined by the value of T , as can be seen by expanding the isospin part of the wave
function,
X
h12|(t1 t2 )T Tz i = ht1 t1z t2 t2z |T Tz ih12|t1 t1z t2 t2z i (6.31)
t1z t2z

With T = 0 and t1 = t2 = t = 1/2, it is

1/2
X
h12|(tt)00 i = htmt − m|00ih1|tmih2|t − mi
m=−1/2
1/2
X (−1)1/2−m (6.32)
= √ h1|tmih2|t − mi
m=−1/2
2
1
1
= √ h12|νπi − h12|πνi = √ h12|νπia
2 2
the isospin part is antisymmetric and therefore the spatial-spin part has to be symmetric which, in turn,
requires J odd.
If T = 1 the isospin part is symmetric and the spatial-spin part is antisymmetric.
6 Two-particle states 68

6.2 Homework problems

Exercise 1: Derive Eqs. (6.7) and (6.8).

Exercise 2: Derive Eq. (6.29). What does it imply?
Chapter 7
Second Quantization

Creation and annihilation operators. Occupation number. Anticommutation relations. Normal product.
Wick’s theorem. One-body operator in second quantization. Hartree-Fock potential. Two-particle Random
Phase Approximation (RPA). Two-particle Tamm-Dankoff Approximation (TDA).

7.1 Creation and annihilation operators

In Fig. 3 of the previous Chapter it is shown the single-particle levels generated by a double-magic core
containing A nucleons. Below the Fermi level (FL) all states hi are occupied and one can not place a
particle there. In other words, the A-particle state |0i, with all levels hi occupied, is the ground state of
the inert (frozen) double magic core.
Above the FL all states are empty and, therefore, a particle can be created in one of the levels denoted
by pi in the Figure. In Second Quantization one introduces the creation operator c†pi such that the state
|pi i in the nucleus containing A+1 nucleons can be written as,

|pi i = c†pi |0i (7.1)

this state has to be normalized, that is,

hpi |pi i = h0|cpi c†pi |0i = h0|cpi |pi i = 1 (7.2)

from where it follows that

|0i = cpi |pi i (7.3)
and the operator cpi can be interpreted as the annihilation operator of a particle in the state |pi i. This
implies that the hole state hi in the (A-1)-nucleon system is

|hi i = chi |0i (7.4)

Occupation number and anticommutation relations

One notices that the number

nj = h0|c†j cj |0i (7.5)

is nj = 1 is j is a hole state and nj = 0 if j is a particle state. In the same fashion it is h0|cj c†j |0i = 1 (0)
if j is a particle (hole) state, that is h0|cj c†j |0i = 1 - nj . Therefore nj is called occupation number of the
state j.
In general it is
h0|c†j ck |0i = nj δjk , (7.6)
and
h0|ck c†j |0i = (1 − nj )δjk . (7.7)
Summing these two equations one gets,

h0|ck c†j + c†j ck |0i = δjk (7.8)

which is valid independently of whether |ji and |ki are particle or hole states. It is a general equation
and, therefore, the creation-annihilation operators satisfy

{cj , c†k } = δjk . (7.9)

7 Second Quantization 70

The operation
{A, B} = AB + BA (7.10)
is called the anticommutator of A and B, and these operators anticommute if

{A, B} = 0 (7.11)

In second quantization the antisymmetrized two-particle state is |ijia = c†i c†j |0i, since it implies

c†i c†j = −c†j c†i =⇒ c†i c†i = 0 (7.12)

as required by the Pauli principle. In the same fashion

ci cj = −cj ci (7.13)

Therefore
{ci , cj } = {c†i , c†j } = 0 (7.14)

Since the state |0i corresponds to a nucleus with A = N + Z nucleons, the state c†i c†j |0i corresponds
to A + 2 nucleons. Therefore
h0|c†i c†j |0i = 0, h0|ci cj |0i = 0 (7.15)
for all i and j.

7.2 Normal product

Since h0|ci c†j |0i = (1 − ni )δij , one can write

ci c†j = (1 − ni )δij + : ci c†j : (7.16)

where : ci c†j : is defined by the relation,

h0| : ci c†j : |0i = 0 (7.17)

The operator : AB : is called normal product between A and B. In the same fashion

c†i cj = ni δij + : c†i cj : (7.18)

From these equations one gets (using {ci , c†j } = δij )

cj c†i = δij − c†i cj = (1 − ni )δij − : c†i cj : (7.19)

but
cj c†i = (1 − ni )δij + : cj c†i : (7.20)
: cj c†i := − : c†i cj : (7.21)
One uses the notation
ci c†j = (1 − ni )δij ; c†i cj = ni δij (7.22)

The operation AB is called contraction of the operators A and B. The contraction is a number. It is
defined as the difference between the ordinary and the normal product of the operators A and B.
Therefore,
c†i c†j = ci cj = 0 (7.23)
from where one gets,
c†i c†j =: c†i c†j :, ci cj =: ci cj : (7.24)
7 Second Quantization 71

Wick’s theorem

One can write any product of creation and annihilation operators in normal form by using the Wick’s
Theorem. It says that the product of operators,

A1 A2 A3 · · · An−1 An (7.25)

where Ai is c†i or ci , can be written as

A1 A2 A3 · · · An−1 An = : A1 A2 A3 · · · An−1 An :
+ A1 A2 : A3 · · · An−1 An :
− A1 A3 : A2 · · · An−1 An :
+ · · · (all single-contractions)
(7.26)
+ A1 A2 A3 A4 : A5 · · · An−1 An :
− A1 A3 A2 A4 : A5 · · · An−1 An :
+ · · · (all double-contractions)
+ · · · (upto n/2-contractions)
The plus or minus sign in each term is determined by the number of permutations one must do in order
to arrive to the final form of the term. An odd (even) number of permutation gives a minus (plus) sign.
The great property of this theorem is that it allows one to get in a straightforward fashion the mean
value of the product of operators, which is what one usually needs. This number is just the term without
normal products, i. e. the last term in the equation above.
We have
h0| : A1 A2 A3 · · · An−1 An : |0i =h0|A1 A2 A3 · · · An−1 An |0i
− h0|A1 A2 A3 · · · An−1 An |0ih0|0i (7.27)
=0
The “normal product” of a normal product

:: AB ::= : AB : + :: AB ::=: AB : (7.28)

The best way of understanding how this theorem works is by applying it to simple cases, We start by
the product
c†i cj = c†i cj + : c†i cj : (7.29)
where only one contraction is possible and no permutation is needed to reach the final value, that is the
sign is plus. As expected, one gets, h0|c†i cj |0i = c†i cj
The next degree of complication is when two contractions are possible, for instance

c†i cj ck c†l = : c†i cj ck c†l :

+ c†i cj : ck c†l : − c†i ck : cj c†l : − cj c†l : c†i ck : + ck c†l : c†i cj : (7.30)

+ c†i cj ck c†l − c†i ck cj c†l

one needs one permutation to get the term c†i ck : cj c†l and therefore a minus sign is added. The same is
done to get the signs of all other terms. The mean value of this operator is,

h0|c†i cj ck c†l |0i = c†i cj ck c†l − c†i ck cj c†l = ni δij (1 − nk )δkl − ni δik (1 − nj )δjl (7.31)

Further examples
We have

: c†i ci : c†α c†β = + ci c†α : c†i c†β : − ci c†β : c†i c†α :

(7.32)
+ : c†i ci c†α c†β :,
7 Second Quantization 72

and

c†α c†β : c†i ci := − c†α ci : c†i c†β : + c†β ci : c†i c†α :

(7.33)
+ : c†α c†β c†i ci :,

from which we get

[: c†i ci :, c†α c†β ] =δiα : c†i c†β : −δiβ : c†i c†α : . (7.34)

We also have
: c†α c†β cγ cδ : c†i c†j = : c†α c†β cγ cδ c†i c†j :

− cγ c†i : c†α c†β cδ c†j : + · · · (other three single contractions) (7.35)

− cγ c†i cδ c†j : c†α c†β : + cγ c†j cδ c†i : c†α c†β :

and one finds, after some algebra

[: c†α c†β cγ cδ :, c†i c†j ] = − δγi : c†α c†β cδ c†j : + · · · (other three single contractions)
(7.36)
+ (1 − ni − nj )(−δγi δδj + δγi δδj ) : c†α c†β :

7.3 One-body operator in second quantization

One-body operators depend upon one radial coordinate r only. In second quantization a one-body
operator M̂ can be written as,
X X
M̂ = hp|M̂ |qic†p cq = hp|M̂ |qi[: c†p cq : + c†p cq ] (7.37)
pq pq

where p and q run over all single-particle states (particle- as well as hole-states). To proof that this is
correct we will evaluate the matrix element of M̂ between two single-particle states, i. e. (A+1)-states
of the form |ii = c†i |0i for which ni = 0. The final result of this calculation should be that we get the
matrix element itself again.
We then evaluate

hi|M̂ |ji =h0|ci M̂ c†j |0i = hp|M̂ |qih0|ci c†p cq c†j |0i
X

pq

hp|M̂ |qih0| ci c†p cq c†j + ci c†j c†p cq |0i

X
=
pq
X h i (7.38)
= hp|M̂ |qi (1 − ni )δip (1 − nj )δqj + (1 − ni )δij np δpq
pq
X
=(1 − ni )(1 − nj )hi|M̂ |ji + (1 − ni )δij np hp|M̂ |pi
p

and we see that with ni = nj = 0 we get the matrix element we needed, i. e. hi|M̂ |ji, but that there is
also another contribution which appears only when i = j. This corresponds to the sum of the mean values
of M̂ over all hole states. It is the interaction of the particles in the A-nucleon core among themselves,
leaving the particle in the (A+1)-nucleus untouched. This term is called ”core polarization”.
To avoid polarization effects one defines
X
M̂ = hp|M̂ |qi : c†p cq : (7.39)
pq

that is, one assumes that M̂ itself includes polarization. One sees that this avoids the core polarization
term, since one cannot contract the indexes p and q (i. e. the term δpq in Eq. (7.38)). Therefore
7 Second Quantization 73

the core polarization effects were assumed to be contained in the operator itself. This procedure is
called ”renormalization”. It is done by introducing some parameters that takes proper account of the
polarization. We will see that in electromagnetic transitions this is done by defining an effective charge
for protons and also for neutrons which, without polarization, has no charge at all.

7.4 Two-body operator in second quantization

To avoid effects related to the interaction of the particles in the core, as it was the core polarization effect
in the one-particle case above, one defines the two-body operator in second quantization in normal form,
i. e. as,
hαβ|M̂ |γδi : c†α c†β cδ cγ :
X
M̂ = (7.40)
αβγδ

and evaluate the matrix element of this operator between antisymmetrized two-particle states, i. e. states
in the (A+2)-nucleus. Our aim is to show that this procedure will indeed provide the antisymmetrized
matrix element. In this context it is worthwhile to point out that a great advantage of second quantization
is that in the many-particle case the Pauli principle is automatically taken into account.
The antisymmetrized two-particle states are,
|ijia = c†i c†j |0i =⇒ a hij| = h0|(c†i c†j )† = h0|cj ci (7.41)
and the matrix element is,

hαβ|M̂ |γδih0|cj ci : c†α c†β cδ cγ : c†k c†l |0i

X
a hij|M̂ |klia = (7.42)
αβγδ

Since the mean value of operators in normal form vanishes, the terms that survive contain only contrac-
tions. They are,
h ih i
cj ci : c†α c†β cδ cγ : c†k c†l = ci c†α cj c†β − ci c†β cj c†α cγ c†k cδ c†l − cγ c†l cδ c†k (7.43)

which give,
X h i
a hij|M̂ |klia = hαβ|M̂ |γδi (1 − ni )δiα (1 − nj )δjβ − (1 − ni )δiβ (1 − nj )δjα
αβγδ
h i
× (1 − nk )δkγ (1 − nl )δlδ − (1 − nk )δkδ (1 − nl )δlγ
h i (7.44)
=(1 − ni )(1 − nj )(1 − nk )(1 − nl ) hij|M̂ |klia − hji|M̂ |klia
h i
− (1 − ni )(1 − nj )(1 − nk )(1 − nl ) hij|M̂ |lkia − hji|M̂ |lkia

In r-representation the matrix element is

Z

∗
hij|M̂ |kli = dr1 dr2 Ψi (r1 )Ψj (r2 ) M̂ (r1 , r2 )Ψk (r1 )Ψl (r2 ) (7.45)

and due to the principle of action and reaction it is, M̂ (r1 , r2 ) = M̂ (r2 , r1 ) which, according to Eq.
(7.45), implies hij|M̂ |kli = hji|M̂ |lki.
The matrix element antisymmetrized to the right only becomes,
   
hji|M̂ |klia = hji|M̂ |kli − |lki = hij|M̂ |lki − |kli = −hij|M̂ |klia (7.46)
and Eq. (7.44) becomes,

a hij|M̂ |klia = (1 − ni )(1 − nj )(1 − nk )(1 − nl ) a hij|M̂ |klia (7.47)

which is just what we wanted to show. However, one does not need to take the matrix element with
antisymmetric wave functions in bra as well as in ket positions since a hij|M̂ |klia = 2hij|M̂ |klia . Therefore
to obtain the matrix element hij|M̂ |klia one has to add a factor 1/4 to the expression (7.40), i. e.
1 X
M̂ = hαβ|M̂ |γδi : c†α c†β cδ cγ : (7.48)
4
αβγδ
7 Second Quantization 74

This is the expression that is used in general. We will use it here also.

7.5 Hartree-Fock potential

We found that to avoid core excitations the one-body operator should be defined in terms of normal
products. That is to use : c†α cβ : instead of c†α cβ . It was due to this that we wrote the two-body operator
in normal form also. But in doing so we bypassed what maybe an important physics. And indeed there
is an important physics behind the core excitations in the case of two-body operators, particularly in the
Hamiltonian. This is what we will explore now.
To this end we write the Hamiltonian H = T + V in a representation consisting of the eigenvectors
of another Hamiltonian. This is often chosen to be an Harmonic oscillator representation because it
is mathematically easy to deal with and also because the nuclear bound states are well described by
Harmonic oscillator potentials, as we have seen in the previous Chapter. Within the chosen representation
(labeled by Greek letters below) the Hamiltonian becomes,
1 X
hαβ|V |γδic†α c†β cδ cγ .
X
H= hα|T |βic†α cβ + (7.49)
4
αβ αβγδ

Converting to normal form one gets,

1 X
hαβ|V |γδi[: c†α c†β cδ cγ :
X
H= hα|T |βi(: c†α cβ : + c†α cβ ) +
4
αβ αβγδ
(7.50)
+ : c†α cγ : c†β cδ − : c†α cδ : c†β cγ − : c†β cγ : c†α cδ + : c†β cδ : c†α cγ

+ c†α cγ c†β cδ − c†α cδ c†β cγ ],

where, as was shown before, we have ci c†j = (1 − ni )δij and c†i cj = ni δij . After some algebra to be
performed,
1 X
H = E0 + HHF + hαβ|V |γδi : c†α c†β cδ cγ : (7.51)
4
αβγδ

where X 1X
E0 = nα hα|T |αi + nα nβ hαβ|V |αβia (7.52)
α
2
αβ

This is the kinetic energy of particles in the occupied states plus the interaction between particles placed
in any pair of levels of the representation. It is the energy carried by the core, as can also be seen by
noticing that E0 = h0|H|0i.
The one-body Hamiltonian is
!
X X
HHF = hα|T |βi + nγ hαγ|V |βγia : c†α cβ : (7.53)
αβ γ

In this Hamiltonian the levels α and β include all states of the representation. These are the levels that
we will occupied by particles which eventually will be added to the core. One thus sees that HHF contains
the core excitations through the interaction of particles in all occupied states (called |γi in HHF ) with
the rest of the particles (including those in the core). The Hamiltonian HHF , which is called the Hartree-
Fock Hamiltonian, thus corresponds to the core excitation which in the one-body case were assumed to
be contained in the renormalized operators.
The diagonalization of HHF provides the Hartree-Fock representation. This is not a very easy task
because it is not a linear problem. To see this we write HHF in Dirac notation, i.e.
!
X X
HHF = |αi hα|T |βi + nγ hαγ|V |βγia hβ| (7.54)
αβ γ
7 Second Quantization 75

and the Hartree-Fock representation will be defined by the eigenvectors {|ii} given by,

HHF |ii = εi |ii (7.55)

To solve this eigenvalue problem we multiply by hα| from the left to get,
!
X X
hα|T |βi + nγ hαγ|V |βγia hβ|ii = εi hα|ii (7.56)
β γ

and the eigenvectors are obtained by imposing the normalization condition,

X
|ii = hα|ii|αi, hi|ii = 1 (7.57)
α

Within the representation {|ii} it should be

hj|HHF |ii = εi δij (7.58)

If it is not then one uses as representation these vectors {|ii} (instead of the one labeled by Greek letters)
to obtained new eigenvectors, which we call {|i0 i}, satisfying
X
HHF |i0 i = εi0 |i0 i, |i0 i = hi|i0 i|ii (7.59)
i

If the condition hj 0 |HHF |i0 i = ε0i δi0 j 0 is still not fulfilled, then one proceeds as before and chooses {|i0 i}
as representation. One repeats this procedure until one arrives after n attempts, to

hj (n) |HHF |i(n) i = εi(n) δi(n) j (n) (7.60)

and the states {|i(n) i} form the Hartree-Fock representation.

This representation is used very often in shell model studies. It has the advantage of being based in
a fully microscopical formalism. This can be contrasted with the one obtained as diagonalization of the
shell model potential discussed in the previous Chapter.

7.6 Two-particle Random Phase Approximation (RPA)

In this Section we will study the dynamics of the (A+2)- and (A-2)-nuclei, that is of two nucleons added
or subtracted from the core. For this we will write the Hamiltonian in the Hartre-Fock representation
which we will label with Greek as well as Latin letters. It is,
1 X
hαβ|V |γδi : c†α c†β cδ cγ :
X
H= εα : c†α cα : + (7.61)
α
4
αβγδ

where εα is the Hartrre-Fock single-particle energy. The constant energy E0 , Eq. (7.52), is not included
because all eigenvalues of the Hamiltonian (7.61) will be referred to the core and, therefore, E0 plays no
role.
To obtain the two-particle energies we evaluate the commutator,
h i X h i 1X h i
H, c†α c†β = εi : c†i ci :, c†α c†β + hij|V |kli : c†i c†j cl ck :, c†α c†β
i
4
ijkl
1
=(εα + εβ )c†α c†β + (1 − nα − nβ ) hij|V |αβia c†i c†j
X
2 (7.62)
ij
1X 1X
− hij|V |βlia : c†i c†j cl c†α : + hij|V |αlia : c†i c†j cl c†β :
2 2
ijl ijl

One sees in this equation that the two-particle creation operators are mixed with three-particle one-hole
excitations, that is with core excitation components. In the Random Phase Approximation (RPA) one
7 Second Quantization 76

neglects the core excitations, that is terms of the form hn2 | : c†i c†j c†α cl : |0i, because they are supposed to
generate states which lie high in the spectrum, thus having little influence over the low-lying two-particle
states. With this and noticing that,

H|n2 i = En2 |n2 i, H|0i = E0 |0i (7.63)

one gets,
h i
hn2 | H, c†α c†β |0i =(En2 − E0 )hn2 |c†α c†β |0i
(7.64)
=(εα + εβ )hn2 |c†α c†β |0i + (1 − nα − nβ ) hij|V |αβia hn2 |c†i c†j |0i
X

i<j

which is the RPA equation. The term 1 − nα − nβ in the RPA equations shows that one can place two
particles above the Fermi level, in which case it is 1 − nα − nβ = 1, or below it (1 − nα − nβ = -1). These
two forms of excitations are mixed to each other, given rise to the so-called RPA correlations. This also
implies that within the RPA one evaluates simultaneously the (A+2)- and (A-2)-systems and, therefore,
there is an influence of one system upon the other.
With ωn2 = En2 − E0 the RPA equation can be written in matrix form as
    
Xn2 A B Xn2
ω n2 = (7.65)
Yn2 −C −D Yn2

where Xn2 (αβ) = hn2 |c†α c†β |0i with α and β particle states and Yn2 (αβ) = hn2 |c†α c†β |0i but with α and β
hole states. In the same fashion the indices of A are all particle states and the indices of D are all hole
states. Instead in the matrices B and C the indices are mixed. For instance C(αβγδ) = hγδ|V |αβia ,
where α and β are hole states while γ and δ are particle states. Notice that the minus sign in front of
the matrices C and D comes from the factor 1 − nα − nβ in Eq. (7.64). Due to this, the RPA matrix
(7.74) is not Hermitian and, therefore, the energies ωn2 can become complex quantities.
Eq. (7.74) can also be written as,
    
Xn2 A B Xn2
ωn2 = (7.66)
Yn2 C D −Yn2

and taking the adjoint of this equation one gets,

 
A B
ωn∗ 2 (Xn∗2 , Yn∗2 ) = (Xn∗2 , −Yn∗2 ) (7.67)
C D

where the properties A† = A, D† = D, C † = B (and, therefore, C = B † ) were used.

Xm2
Multiplying Eq. (7.67) to the right by and using Eq. 7.66 one gets,
−Ym2
 
∗ ∗ ∗ Xm2
(ωn2 − ωm2 )(Xn2 , Yn2 ) =0 (7.68)
−Ym2

If n2 6= m2 this Equation implies that

 
Xm2
(Xn∗2 , Yn∗2 ) =0 (7.69)
−Ym2

and normalizing it to unity one gets,

(1 − nα − nβ )hn2 |c†α c†β |0i∗ hm2 |c†α c†β |0i = δn2 m2

X
(7.70)
α≤β

which defines the RPA scalar product, or metric. In functional analysis it is also called indefinite inner
product.
The two-particle state can be written as,

X(αβ, n2 )c†α c†β |0i

X
|n2 i = (7.71)
α≤β
7 Second Quantization 77

and multiplying by hm2 | one gets

X(αβ, n2 )hm2 |c†α c†β |0i

X
δ n2 m 2 = (7.72)
α≤β

since the basis elements form an independent set one finds, comparing with Eq. (7.70),

X(αβ, n2 ) = (1 − nα − nβ )hn2 |c†α c†β |0i∗ (7.73)

which is the RPA wave function amplitude.
As an example, we evaluate the energy of the a system with one particle orbital (α) and one hole
orbital (β). The RPA equation can be written in a 2 × 2 matrix form as
    
x A B x
ωn2 = . (7.74)
y −C −D y

We have A = 2εα + hαα|V |ααiJT where JT denote the spin and isospin of the two particle state.
D = −2εβ + hββ|V |ββiJT .q B = C = hαα|V |ββiJT . In such a case we get two solutions for the
A−D 2
eigen values, ωn2 = A−D


2 ± 2 + AD − C 2 , corresponding to the particle-particle and hole-hole
excitations respectively. For a system with no particle-hole correlation, we get ωn2 = A for the two-
particle state and ωn2 = −D for the two-hole state.

7.7 Tamm-Dankoff Approximation (TDA)

We will concentrate in the shell model in this course, and here one has either two-particle or two-hole
excitations, and the (A+2) and (A-2) systems are independent of each other. The shell model cases are
actually particular cases of the RPA since one gets them by imposing the condition that only particles
can occupied particle states and only holes can occupied hole states. This is called Tamm-Dankoff
approximation (TDA).
This approximation implies that the matrices B and C vanish in Eq. (7.74). The particle- and
hole-states decoupled and the RPA equation transforms in two TDA equations, one for particle states, i.
e.
ωn2 Xn2 = AXn2 (7.75)
and the other one for hole states,
−ωn2 Yn2 = DYn2 (7.76)
Since the matrices A and D are Hermitian the energies are real, as they should be.
We will study these two cases separately starting from Eq. (7.64). For the two-particle case the TDA
thus means nα = nβ = 0. That is
h i
hn2 | H, c†α c†β |0i =(En2 − E0 )hn2 |c†α c†β |0i
(7.77)
=(εα + εβ )hn2 |c†α c†β |0i + hij|V |αβia hn2 |c†i c†j |0i
X

i<j

which is the TDA equation. It is also the shell model equation, which we will apply in the next Chapter.
The equation for the two-hole states is better obtained starting from the transverse of the RPA
h i†
operator form, i. e. by performing the operation H, c†α c†β = (Hc†α c†β )† − (c†α c†β H)† =cβ cα H − Hcβ cα
= [H, cα cβ ] in Eq. (7.62). One thus obtains,
X
[H, cα cβ ] = − (εα + εβ )cα cβ − (1 − nα − nβ ) hij|V |αβia ci cj (7.78)
i<j

where the contribution from three-particle one-hole operators have been neglected. Since the single-
particle levels are hole states, one has (1 − nα − nβ ) = -1 and the TDA equation for the states |n2 i in
the (A-2) nucleus is
hn2 | [H, cα cβ ] |0i =(En2 − E0 )hn2 |cα cβ |0i
= − (εα + εβ )hn2 |c†α c†β |0i + (7.79)
X
hij|V |αβia hn2 |ci cj |0i
i<j
7 Second Quantization 78

This TDA equation is also the shell model equation for two-hole states. It is usually written as,

ω(n2 )hn2 |cα cβ |0i

=(εα + εβ )hn2 |c†α c†β |0i − (7.80)
X
hij|V |αβia hn2 |ci cj |0i
i<j

and the eigenvalues ω(n2 ) = −(En2 − E0 ) are minus the energies of the (A-2)-nucleus referred to the
ground state of the A-nucleus, i. e. to the core. Remember that for convention these energies are minus
the binding energies (with this convention the binding energies are all positive). The binding energy of a
nucleus increases with the nuclear number A. Therefore En2 − E0 > 0 and ω(n2 ) < 0. The eigenvalues of
the TDA equation for the (A+2) nucleons, Eq. (7.77), are also negative, as it should be for bound states.
The difference between particles and holes is that for holes the interaction contributes with a minus sign,
as seen in Eq. (7.80).
The TDA wave function can be written in the two-particle basis {c†α c†β |0i}, where it should be α < β
because the states αβ and βα are related by {c†α c†β |0i} = -{c†β c†α |0i}. One thus gets,

X(αβ; n2 )c†α c†β |0i

X
|n2 i = (7.81)
α<β

The TDA eigenvectors hn2 |c†α c†β |0i and the wave function amplitudes X are related by,

X(αβ; n2 )hm2 |c†α c†β |0i

X
hm2 |n2 i = δm2 n2 = (7.82)
α<β

since the basis states c†α c†β |0i form an independent set of unit vectors, it should be X(αβ; n2 ) = hn2 |c†α c†β |0i∗ .
For a system with two particles in one orbital α, we simply have ωn2 = 2εα + hαα|V |ααiJT where JT
denote the spin and isospin of the two particle state.
7 Second Quantization 79

7.8 Homework problems

Exercise 1:
a) Write the operator c+
p cq in normal form.
b) Write the operator c+ +
p cq cr cs in normal form.

Exercise 2:
1 +
a) Write the Hamiltonian H = αβ < α|T |β > c+ < αβ|V |γδ > c+
P P
α cβ + 4 αβγδ α cβ cδ cγ in normal
form and extract the Hartree-Fock potential.
b) Evaluate < 0|H|0 >.

Exercise 3:
+ +
Write the operator : c+ +
p cq cr cs :: ci cj : in normal form.

Exercise 4:
in the Hartree-Fock representation, i. e. H = α α : c+
P
a) With
P the Hamiltonian H+ written α cα :
+(1/2) αβγδ < αβ|V |γδ >: cα c+ + +
β cδ cγ : evaluate the commutator [H, cα cβ ] in normal form and extract
the two-particle RPA equation.
b) Starting from the RPA metric deduce the metric of the TDA space.
Chapter 8
Shell model excitations

Two-particle energies and wave functions. Interaction matrix elements. The separable matrix element.
Pairing force.

8.1 Two-particle energies and wave functions

We have seen that the second quantization formalism is very appropriate to treat a many-nucleon system
because it takes into account the Pauli principle in a straightforward fashion. Within this formalism
we have derived the two-particle shell model (TDA) equations, which we will apply in this Chapter.
As discussed in the previous Chapter, the first step in the solution of a many-body problem in nuclei
is to choose a representation. Within the shell model the vectors {|αi} in the representation are the
eigenvectors of the free shell model Hamiltonian H0 , i. e. H0 |αi = εα |αi. The core consists of a double
magic nucleus with A nucleons in its ground state with energy E0 (A). The single-particle energy εα is
the energy of the A + 1 nucleus referred to the core, i. e. εα = Eα (A + 1) − E0 (A). The energies of
the A + 2 nucleus are En2 (A + 2). All energies, i. e. E0 (A), Eα (A + 1) and En2 (A + 2) are minus the
binding energies of the nuclei in their respective states. Notice that if |n2 i is an excited state of the
(A+2)-nucleus, then the binding energy is the binding energy of the ground state minus the energy of
the excited state |n2 i (since the excited state is less bound than the ground state).
The TDA equation provides the energies of the A + 2 nucleus relative to the core in terms of the
interaction matrix elements as,

(En2 − E0 )hn2 |c†α c†β |0i = (εα + εβ )hn2 |c†α c†β |0i + hγδ|V |αβia hn2 |c†γ c†δ |0i
X
(8.1)
γ<δ

The two-particle wave functions is

X(γδ; n2 )c†γ c†δ |0i

X
|n2 i = (8.2)
γ≤δ

where the indices are all particle states.

With ωn2 = En2 − E0 the shel model systems of equations becomes,
X 
(−ωn2 + εα + εβ )δαγ δβδ + hγδ|V |αβia X(γδ; n2 ) = 0 (8.3)
γ<δ

where we have replaced X(γδ; n2 ) = hn2 |c†γ c†δ |0i, and not its complex conjugate, because we will take the
interaction matrix elements as real.
The system of equations (8.3) has no trivial solution only if the determinant vanishes, i. e.

||(εα + εβ )δαγ δβδ + hγδ|V |αβia | = 0 (8.4)

If there are M elements in the representation, then one gets M energies ωn2 from this condition, and one
of the equations in Eq. (8.3) becomes a linear combination of the others. One thus disregards one of the
equations and, instead, uses the normalization condition
X
X 2 (γδ; n2 ) = 1 (8.5)
γ≤δ

We will apply the two-particle shell model equations in an illustrative example. For this purpose
we consider the case of two protons outside the core 8838 Sr50 . The number Z = 38 can be considered a
semi-magic number, since for the low lying states the effects of the shells below Z = 38 are not important.
8 Shell model excitations 81

The neutrons and protons fills different shells and the isospin formalism is not well suited in this case.
We will consider neutrons and protons as different particles. The neutrons, filling the shell N = 50,
are frozen and do not contribute to the two-proton spectrum. The single-particle states are then those
proton shells between Z = 38 and Z = 50, which are p1/2 and g9/2 . We will designate the corresponding
single-particle energies as ε1 = εp1/2 and ε2 = εg9/2 .
The possible two-proton states are

(p1/2 )2 −→ 0+
(p1/2 g9/2 ) −→ 4− , 5− (8.6)
(g9/2 )2 −→ 0+ , 2+ , 4+ , 6+ , 8+

Which shows that in this case there is only one basis state, except for 0+ for which there are two. We
will start analyzing these 0+ states.
With |αi = |(p1/2 )2 ; 0+ i and |βi = |(g9/2 )2 ; 0+ i the shell model equations read,

(2ε1 + hα|V |αi − ωn )X(α; n) + hα|V |βiX(β; n) = 0

(8.7)
hβ|V |αiX(α; n) + (2ε2 + hβ|V |βi − ωn )X(β; n) = 0

which provides the determinant,

2ε1 + hα|V |αi − ωn hα|V |βi

=0 (8.8)
hβ|V |αi 2ε2 + hβ|V |βi − ωn

from where the energies are obtained, i. e.

" 2 #1/2
Vαα + Vββ Vαα − Vββ 2
ωn = ε1 + ε2 + ± ε1 − ε2 + + Vαβ (8.9)
2 2

where the notation Vγδ = hγ|V |δi was used.

With these values of the energies the two shell model Eqs. (8.7) are linearly dependent upon each
other. One obtains the wave function amplitudes from one of them as,

2ε1 + hα|V |αi − ωn

X(β; n) = − X(α; n) (8.10)
hα|V |βi

and the other is the normalization condition which gives,

2ε1 + hα|V |αi − ωn 2

X 2 (α; n) + X 2 (β; n) = X 2 (α; n) 1 + ( ) =1 (8.11)
hα|V |βi

which provides X(α; n) and from Eq. (8.10) one obtains X(β; n).
For the other states there is only one configuration which we will call |αi. Therefore the wave function
amplitude is X(α; n) = 1 and the energy of the state is, ωn = 2ε + Vαα
In the next Section we will outline procedures to evaluate the interaction matrix elements.

8.2 Interaction matrix elements

One important assumption of the shell model is that it exists a frozen core and that the nucleons moving
outside the core feels the interaction of the core as a mean field. This is the shell model potential that
we have described in Chapter 4. The dynamics of the nucleus is then determined by the particles moving
outside the core. These are called valence particles.
An accurate knowledge of the interaction among the valence particles is of a fundamental importance
for a proper description of the nuclear spectra. This is not an easy task since the interaction among
nucleons inside the nucleus is very different than the one corresponding to free nucleons. The core
induces the central field that generates the shell model representation. But the interaction among valence
particles is also affected by the core. Due to this one uses effective forces which somehow take into
account the influence of the nucleons in the core upon the valence particles. A sensible way to include
the influence of the core upon the interacting particles is to determine the interaction matrix elements
8 Shell model excitations 82

from the experimental data of the two-nucleon system. That is, the two-nucleon shell model equations
(8.3) can be written as
X
[ωn2 − (εα + εβ )]X(αβ; n2 ) = hγδ|V |αβia X(γδ; n2 ) (8.12)
γ<δ
P
and since n2 |n2 ihn2 | = 1 one gets,
X
hγδ|V |αβia = X(αβ; n2 )[ωn2 − (εα + εβ )]X(γδ; n2 ) (8.13)
n2

where we have assumed that the matrix elements, and therefore the amplitudes X, are real numbers. This
expression shows that one can extract the interaction matrix elements from experiment if the amplitudes
as well as the single-particle and the two-particle energies are measured. This happens in some cases,
specially when the dimension of the shell model space is small. The 88 Sr example given above is one of
those cases. But in general what one does is to extract as many matrix elements from experiment as
possible and to determine the others by variations around reasonable values. One gets the final matrix
elements when the spectra of the nuclei under study are fairly reproduced. This is a procedure been
used at present. With the relatively few parameters that includes such fitting process (all corresponding
to the two-nucleon system) one can study the dynamics of a nucleus with many valence particles. The
dimensions of the corresponding shell model matrices can reach huge values, like 1010 and even more.
The ”reasonable” values of the interaction matrix elements mentioned above are often taken to be
those provided by well probed two-nucleon potentials. Some of the simplest of those potentials, namely
the separable potential and the pairing potential, are described below.

The separable matrix element

One of the interaction matrix element which is often used in nuclear physics is taken to be separable.
This is defined, for states of spin J, as

hpq; J|V |rs; Ji = −Gf (pq; J)f (rs; J) (8.14)

where G is a constant called ”strength” and f (pq; J) is a function of the states p and q. One uses
various expressions for this function, although an important criterion to judge its quality is how well it
fits experimental data.
With the separable matrix element the shell model equations become,
Xh i
(εp + εq − ωn )δpr δqs − Gf (pq; J)f (rs; J) X(rs; n) = 0 (8.15)
r6s

which can be written as,

f (pq; J) X
X(pq; n) = G f (rs; J)X(rs; n) (8.16)
εp + εq − ωn
r6s

X
multiplying this equation by f (pq; J) one gets
p6q

X X f 2 (pq; J) X
f (pq; J)X(pq; n) = G f (rs; J)X(rs; n) (8.17)
εp + εq − ωn
p6q p6q r6s

that is,
X f 2 (pq; J)
G =1 (8.18)
ε p + ε q − ωn
p6q

This equation is called ”Dispersion Relation”. To get the energies from the dispersion relation one has
first to determine the strength G. This is extracted from the dispersion relation by using for ωn the
experimental value of the energy of the lowest state with spin J (this is called ”yrast” state), which for
8 Shell model excitations 83

the spin 0+ is the ground state. The energies of the other states with spin J are those that satisfies the
dispersion relation.
The wave function amplitudes are, according to Eq. (8.16), given by
f (pq; J)
X(pq; n) = Nn (8.19)
εp + εq − ωn
where Nn is the normalization constant, i. e.
X f 2 (pq; J) −1/2
Nn = (8.20)
εp + εq − ωn
p6q

Pairing force

In nuclear physics the pairing force is a particular case of the separable force. It is generally applied to
study the states 0+ . It is assumed that the orbitals p and q are time reversal states of each other. This
implies that in a coupled-scheme the two states are exactly the same. The corresponding value of the
function f is,
f (pp; 0+ ) = 2jp + 1
p
(8.21)
To illustrate how the separable force in general, and the pairing force in particular, is applied we will
briefly present the states in 90 Zr studied above. In this case the functions f acquire the values,
√ √
f (α; 0+ ) = f (p21/2 ; 0+ ) = 2; f (β; 0+ ) = f (g9/2
2
; 0+ ) = 10 (8.22)
from where one evaluates the strength as,
!−1
2 10
G= + (8.23)
2ε1 − E0+ 2ε2 − E0+
1 1

and the wave function amplitudes are,

√ √
2 10
X(α; n) = Nn , X(β; n) = Nn (8.24)
2ε1 − ωn 2ε2 − ωn
The numerical values of the energies are obtained by solving Eq. (8.18). With these energies one evaluates
the wave function amplitudes with the normalization calculated as in Eq. (8.20).

8.3 Isospin symmetry in nuclei

As an illustrative example we will analyze the case of two nucleons outside the core 168 O8 (Oxygen 16).
The total isospin of the two-nucleon systems can be T =0 or T =1. For the nuclei with Tz = ±1 it must
be T ≥ 1. This is the case in 18 18 18
10 N e8 (Neon 18) and 8 O10 . In 8 F9 (Fluor 18) it is T ≥ 0.

The single-particle states above the magic number 8 are 0d5/2 , 1s1/2 and 0d3/2 . However the lowest
two-nucleon states should be mostly due to the lowest shell, i. e. 0d5/2 . Therefore the lowest two nucleons
states have to be 0+ , 2+ and 4+ if T = 1 and 1+ , 3+ and 5+ if T = 0.
As seen in Fig. 8.1 one indeed finds that the three lowest levels are 0+ , 2+ and 4+ for the two-proton
18
10 N e8 and the two-neutron 188 O10 nuclei, for which it has to be T = 1. But even the T = 1, Tz = 0 state
0+ is seen in 18 F , although here the corresponding 2+ and 4+ states have not been detected yet. Instead
in this T = 0 nucleus the odd states 1+ , 3+ and 5+ appear.
We notice that the relative energies of the T = 1 states, i. e. the states in 18 Ne and 18 O, should not
be influenced by the Coulomb energy since this affects equally to all levels of a given nucleus, and the
energy difference cancels those effects. This explains why the excitation energy of the states 2+ are very
similar in both nuclei, although this is not the case for the state 4+ . These differences between theory
and experiment in the isoscalar as well as in the isovector excitations are due to the shells 1s1/2 and 0d3/2
which we neglected. Their influence is more important as the energy of the excited level increases.
A detailed calculation of these levels requires the inclusion of many shells as well as a proper knowledge
of the interaction matrix elements. We will do this within the shell model (TDA) formalism that we
developed in the previous Chapter.
8 Shell model excitations 84

+
+ 3 .5 5 5 4
3 .3 7 6 4

-
+ 2 .1 0 1 2 1 .9 8 2 2
+
1 .8 8 7 2 +
1 .7 0 1 1
+
1 .1 2 1 5 +
1 .0 4 2 0
+
0 .9 3 7 3

+ + +
0 0 0 0 0 0
1 8 1 8 1 8
1 0
N e 8 9
F 9 8
O 1 0

18 18 18
Figure 8.1: Nuclear spectra of the nuclei 10 O8 , 9 F9 and 8 O10 .

NN NP PP

Figure 8.2: The spectra of mirror nuclei 54 Ni, 54 Co and 54 Fe (with T = 1). From PRL 97, 152501 (2006).
8 Shell model excitations 85

48 48
Figure 8.3: The spectra of mirror nuclei Mn and V (with T = 3/2). From PRL 97, 132501 (2006).

8.4 Nucleon-Nucleon Interaction and LS coupling

If one is interested in the low-energy region where the nucleons hardly get excited internally, we can treat
the nucleons as inert, structureless elementary particles, and we can understand many of the properties
of the multi-nucleon systems by the nucleon-nucleon interactions. If the nucleons are non-relativistic,
the interaction can be described by a potential. Since the fundamental theory governing the nucleon-
nucleon interactions is QCD, the interactions shall be calculable from the physics of quarks and gluons.
Nonetheless, the problem is quite complicated and only limited progress has been made from the first
principles so far. Therefore, the approach we are going take is a phenomenological one: One first tries
to extract the nucleon- nucleon interaction from the nucleon-nucleon scattering data or few nucleon
properties, and then one tries to use these interactions to make predictions for the nuclear many-body
system.

General Aspects of the Nucleon-Nucleon interaction

For the two-nucleon system, the experimental information consists of two particle scattering phase shifts
in various partial waves, and the bound state properties of the deuteron. From data, one learns that the
nuclear force has the following the main features:
• Attractive: to form nuclear bound states
• Short Range: of order 1 fm

• Spin-Dependent
• Noncentral: there is a tensor component
• Isospin Symmetric
• Hard Core: so that the nuclear matter does not collapse

• Spin-Orbit Force
• Parity Conservation
8 Shell model excitations 86

In phenomenology, one tries to come up with the forms of the forces which will satisfy the above properties.
In particular, one parametrize the short distance potentials consistent with fundamental symmetries and
fit the parameters to experimental data.
If we consider the nucleons as elementary particles, their strong interactions are known have a short
range from α particle scattering experiments of Rutherford. In fact, the range of the interaction is roughly
the size of the atomic nuclei, namely on the order of a few fermis (fm) (1fm = 10−15 m). The first theory
of the nucleon force was put forward by H. Yukawa, who suggested that the interaction between two
nucleons is effected by the exchange of a particle, just like the interaction between the electric charges
by the exchange of a photon. However, because the nucleon interactions appear to be short-ranged, the
particle must have a finite mass. In fact, one can correlate the range and mass roughly by the quantum
uncertainty principle
r ∼ 1/m (8.25)
therefore, the mass of the quanta exchanged is about 1/fm which is about 200 MeV. The particle was
discovered almost 20 years later and was identified as π meson (140 MeV). The most significant aspect
of the Yukawa theory is generalizing the relation between particles and forces – the existence of strong
interactions implies the existence of a new particle! This was considered a novel and radical idea at that
time.
The modern theory of interactions through particle exchanges is made possible by the development
of quantum field theory. However, at low-energy, one can assume the interactions is instantaneous and
therefore the concept of interaction potential becomes useful.

Yukawa theory and One-Pion Exchange potential

To illustrate the derivation of a potential through particle exchange, consider the interaction of a scalar
particle with the nucleons. Introducing a quantum field φ to represent the particle, the Yukawa interaction
has the form,
g 2 e−mr
V (r) = − (8.26)
4π r
which is an attractive interaction, independent of the spin of the nucleon. When m = 0, the interaction is
Coulomb-like and is long range. For a finite m, the interaction is negligible beyond the distance r ∼ 1/m.
Pi-meson is the lightest meson because of the spontaneous chiral symmetry breaking. It is responsible
to the longest range nucleon-nucleon force. One of the important new features here is isospin symmetry.
Pi-meson has iso-spin 1, and can be represented by a vector π = (π1 , π2 , π3 ) in isospin space. Combi-
nations of π1,2 generate positive and negative charged pions π ± , and π3 corresponds to a neutral pion
π0 .
Because the pion is a pseudo-scalar particle, parity conservation requires that it be emitted or absorbed
in the p-wave. Therefore there is a factor of p associated with the interaction vertex in the momentum
space. This factor is multiplied by the spin operator σ forms a pseudo-scalar. Putting together all the
factors, one finds that the nucleon-nucleon potential from one pion exchange as,

g2 e−mπ r
V (r) = 2 (τ 1 · τ 2 )(σ1 · ∇)(σ2 · ∇) (8.27)
4MN r
where the isospin factor τ1 · τ2 depends on the isospin states of the two nucleons. This potential matches
the phenomenological forms extracted from experimental data at large N-N separations (∼ 2 − 3 fm).
It is therefore advantageous to form states of good isospin to describe a system of 2 nucleons. Note the
combination τ1 · τ2 is given by
  
3 −3, T = 0
τ1 · τ2 = 2 T (T + 1) − = (8.28)
2 1, T = 1

At smaller distance, there is also exchanges from scalar meson (isospin 0) of about 500 MeV. The
interaction is attractive as we seen above, corresponding to a medium range attraction. Finally there are
also exchanges from vector mesons, ω (isospin-0) meson and ρ meson (spin-1). The interactions from the
ρ and ω meson are short-range repulsive. One can build a phenomenological nucleon-nucleon interaction
based on the meson exchanges. The so-called Nijmegen potential and Bonn potential are generated
through this approach. The short range interactions are model dependent by nature, there is no unique
picture for them. This is so because only in the low-energy processes, potentials are useful concept,
however, the short range interactions are not so sensitive to the low-energy observables. Therefore, either
8 Shell model excitations 87

one can adopt a more phenomenological approach or use the so-called effective field theory approach to
parameterize the unknowns systematically. During the 1990s several groups have constructed socalled
high-precision, charge-dependent NN potentials (CD-Bonn, AV18, etc).
A brief history of NN interactions
• 1935 Yukawa (meson theory or Meson Hypothesis)
• 1950s Full One-Pion-Exchange potential (OPEP): Hamada-Jonston
• 1960s non-relativistic One-Boson-Exchange potential (OBEP) (pions, Many pions, scalar mesons,
782(ω), 770(ρ), 600(σ; not observed yet))
• 1970s fully relativistic OBEPs
– 2-pion exchange
– Paris, Bonn potential
• 1990s High-precision Nijmegen, Argonne V18, Reid93, Bonn potentials
• 1990-2000s Chiral or Effective Field Theory potentials (2 and 3 body), Lattice QCD
Comonly used simple NN potentials
• Square well V = −V0 (0) for r ≤ R(r > R) where R ∼ 2fm.
2 2
• Gaussian: V = −V0 e−(µr) = −V0 e−(r/R) where R ∼ 1.5fm.
• Exponential: V = −V0 e−µr = −V0 e−r/R where R ∼ 0.75fm.
• Yukawa: V = −V0 e−µr /µr = −V0 e−r/R /(r/R) where R ∼ 2.5fm.
• One-boson-exchange potentials: V = −V0 (1 + 1/µr)e−µr /µr, V = −V0 (1 + 3/µr + 3/(µr)2 )e−µr /µr

LS coupling
For the system of two nucleons, use r = r1 − r2 to represent relative position and p = (p1 − p2 )/2 relative
momentum, s1 and s2 their respective spins. The relative orbital angular momentum is L = r × p and
the total spin is S = s1 + s2 . When the spins are coupled, the total spin can either be 0 or 1. For the
case of S = 0, we have a single spin state which is called singlet. For the case of S = 1, we have three spin
states which are called triplet. The total angular momentum is the sum of orbital angular momentum
and total spin: J = L + S. In the singlet spin case, we have J = L because S = 0. For the triplet states,
J = L − 1, L, L + 1 if L 6= 0, and J = 0 if L = 0. A state with (S, L, J) is usually labeled as 2S+1 LJ
, where L = 0, 1, 2, 3, are usually called S, P, D, F, G... states. We choose the basis states with good J,
|n(LS)JM i where, X
|n(LS)JM i = |YLmL i|(s1 , s2 )SmS ihLmSmS |JM i (8.29)
mL
M
where hLmSmS |JM i is the Clebsch-Gordon coefficient. In the coordinate representation, we use YLSJ
to label the above total angular momentum eigenstates. Therefore, the eigenfunction can be written as
M
ψnLSJM = Rn(SL)J (r)YLSJ (8.30)
where Rn(SL)J (r) is the corresponding radial wave function.
For the two nucleon system, the spin S and isospin T can either be 1 or 0. Isospin symmetry requires
that the wave function reverse its-sign upon an odd permutation of alI coordinates (i.e. space, spin and
isospin) of any two nucleons. Since the total wave function has to be antisymmetric, let us consider what
are the possible value of T , S, and L to make that happen. If we use +1 to represent symmetric wave
function and −1 to represent antisymmetric, then the spin wave function is (−1)S+1 and the isospin wave
function has symmetry factor (−1)T +1 . The orbital wave function is (−1)L (i.e., r → −r). The total
symmetry factor is (−1)L+S+T which has to be −1. Therefore L + S + T has to be odd. We have
h12|ψnLSJM T Tz ia = Rn(SL)J (r)(h12|YLmL (r)ih12|(1/21/2)SSz ih12|(1/21/2)T Tz i)a (8.31)
and
h12|ψnLSJM ia = N (1 − (−1)L+S+T )Rn(SL)J (r)h12|YLmL (r)ih12|(1/21/2)SSz ih12|(1/21/2)T Tz i (8.32)
8 Shell model excitations 88

Table 8.1: The four LST combinations that are compatible with the Pauli principle.
φ(L) χ(S) ζ(T)
even+ singlet(S=0)− triplet(even,T=1)+
even+ triplet(S=1)+ singlet(odd,T=0)−
odd− singlet(S=0)− singlet(odd,T=0)−
odd− triplet(S=1)+ triplet(even,T=1)+

LS coupling in laboratory frame

One may also rewritten the wave functions of Eqs. (6.6) & (6.7) in term of LS coupling. In the case of
two nucleon in lp we have

h12|ψ(lp ,lp )LSJM ia =N (1 − (−1)L+S+T )Rp (r1 )Rp (r2 )

(8.33)
× h12|(lp lp )LM ih12|(1/21/2)SSz ih12|(1/21/2)T Tz i.

In general we have

h12|ψ(lp ,lq )LSJM ia =N [Rp (r1 )Rq (r2 )h12|(lp lq )LM ih12|(1/21/2)SSz ih12|(1/21/2)T Tz i
− (−1)lp +lp +L+S+T )Rq (r1 )Rp (r2 ) (8.34)
×h12|(lq lp )LM ih12|(1/21/2)SSz ih12|(1/21/2)T Tz i]

The parity of the system is determined by −1lp +lq . In the relative frame it is −1L .

Tensor force
I. Central force
What are the possible forms of the nuclear force? The simplest is a pure central force which just
dependent on the relative distance VC (r). In this case, different L states have different energies. The
eigen-functions of the system can chosen to be |nLmL SmS i ∼ RnL (r)YLmL χSmS .
There could be also a pure spin-dependence force. The most general form is VS (r)σ1 · σ2 . In fact,
we can write
σ1 · σ2 = 2S 2 − 3 (8.35)
Therefore the matrix element of the spin operator depends on the total spin of the two particles. In the
singlet state, we have σ1 · σ2 = −3, the potential is VC − 3VS ; in the triplet state (S = 1), we have
σ1 · σ2 = 1, and the potential is VC + VS . Now, the energy not only depends on the quantum number L
but also on S. However, the eigen-function of the system can still be chosen as |nLmL SmS i; the radial
wave function depends on S, RnLS (r), because the potential does.
There can be also a pure iso-spin-dependence force. The most general form is VI (r)τ1 · τ2 ; There can
be a spin-isospin dependent force. The most general form is VSI (r)σ1 · σ2 τ1 · τ2 .

II. Spin-orbit force

The spin-orbit force can be written as VLS (r)L · S = 21 VLS (r)L · (σ1 + σ2 ). We may also have a
spin-orbit-isospin dependent force VLSI (r)L · S = 21 VLSI (r)L · (σ1 + σ2 )(τ1 · τ2 ).
The matrix element of the operator would be simple in the basis which is formed by the common
eigenstates of J 2 , J3 , L2 , S 2 and T 2 . The potential used for solving the radial Schroedinger equation is

VLSJT (r) =VC (r) + [2S(S + 1) − 3]VS (r) + [2T (T + 1) − 3]VI (r)
+ [2S(S + 1) − 3][2T (T + 1) − 3]VSI (r)
1 (8.36)
+ [J(J + 1) − L(L + 1) − S(S + 1)]VLS (r)
2
1
+ [J(J + 1) − L(L + 1) − S(S + 1)][2T (T + 1) − 3]VLS (r)
2
8 Shell model excitations 89

Another useful way is to write the potential in the singlet(S = 0)-triplet(S = 1)-odd(T = 0)-even(T =
1) representation. For the central force we have four components
1 1
VCSO (r) =UCSO (r) (1 − σ1 · σ2 ) (1 − τ1 · τ2 ) = VC (r) − 3VS (r) − 3VI (r) + 9VSI (r)
4 4
SE SE 1 1
VC (r) =UC (r) (1 − σ1 · σ2 ) (3 + τ1 · τ2 ) = VC (r) − 3VS (r) + VI (r) − 3VSI (r)
4 4 (8.37)
TE TE 1 1
VC (r) =UC (r) (3 + σ1 · σ2 ) (1 − τ1 · τ2 ) = VC (r) + VS (r) − 3VI (r) − 3VSI (r)
4 4
TO TO 1 1
VC (r) =UC (r) (3 + σ1 · σ2 ) (3 + τ1 · τ2 ) = VC (r) + 3VS (r) + VI (r) + VSI (r)
4 4
The spin-orbit force has two components for the isovector (even) and isoscalar (odd) channels
1 1
VLSE (r) =ULSE (r) L · (σ1 + σ2 ) (1 − τ1 · τ2 )
2 4 (8.38)
1 1
VLSO (r) =ULSO (r) L · (σ1 + σ2 ) (3 + τ1 · τ2 )
2 4
They only work on the S = 1 channel.

II. Tensor force

Finally, there is an interaction of the type
 
(σ1 · r)(σ2 · r)
VT (r) 3 − σ1 · σ2 (8.39)
r2

which is called the tensor interaction and the associated structure is often denoted as S12 . As before,
one may also have a tensor-isospin dependent force VT I . Tn the singlet-triplet-odd-even representation,
the tensor force also has two components for the isovector (even) and isoscalar (odd) channels
 
(σ1 · r)(σ2 · r) 1
VT N E (r) =UT N E (r) 3 2
− σ1 · σ2 (1 − τ1 · τ2 )
r 4
  (8.40)
(σ1 · r)(σ2 · r) 1
VT N O (r) =UT N O (r) 3 − σ1 · σ2 (3 + τ1 · τ2 )
r2 4

The following property of the Pauli matrics is useful

(σ · A)(σ · B) = (A · B) + iσ · (A × B), (8.41)

from which we have

(σ · r)2 = (r · r) + iσ · (r × r) = r2 . (8.42)
Thus one has
2(σ1 · r)(σ2 · r) =(σ1 · r + σ2 · r)2 − (σ1 · r)2 − (σ2 · r)2
(8.43)
=(σ1 · r + σ2 · r)2 − 2r2 .

Using the total spin operator, we can write

(σ1 · r)(σ2 · r) = 2(S · r)2 − r2 , (8.44)

Then the tensor structure becomes

(S · r)2
 
2
S12 = 2 3 −S . (8.45)
r2
2
h i
(σ1 ·r)(σ2 ·r)
It can be easily show that for Q = (S·r)
r 2 = 1
2 1 + r 2 we have

Q2 = Q. (8.46)
8 Shell model excitations 90

The action of Q̂ on the total angular momentum eigenstates are then

M
QYJ0J = 0
M M
QYJ1J = YJ1J
M M
p M
(2J + 1)QYJ+11J = JYJ+11J + J(J + 1)YJ−11J
M
p M M
(2J + 1)QYJ−11J = J(J + 1)YJ+11J + (J + 1)YJ−11J . (8.47)

In the presence of the tensor interaction, L2 and S 2 no longer commute with the Hamiltonian. The
states with the same J but different L can mix under the interaction. The states can mix only when
their orbital angular momenta differ by 2 unit.
Going back to the one-pion exchange potential, it can be shown that

e−mπ r
   −mπ r
1 3 3 e
(σ1 · ∇)(σ2 · ∇) = m2π σ1 · σ2 + S12 1 + + 2 2 . (8.48)
r 3 mπ r mπ r r
8 Shell model excitations 91

8.5 Homework problems

Exercise 1:
Taking the nucleus 40 Ca as a core,
a) Determine the energy of the neutron state 0f7/2
2 2
b) Determine the values of the matrix elements < f7/2 ; λ|V |f7/2 ; λ >, with 0 ≤ λ ≤ 6, that adjust the
42
experimental spectrum in Ca.
Hint: Assume that the last neutrons only occupy the neutron orbital 0f7/2 for low-lying states in
42
Ca.

Exercise 2 (see Section 8.1): p

a) Assume 88 Sr as the core and the pairing interaction hj 2 ; J = 0|V |j 02 ; J = 0i = −G (2j + 1)(2j 0 + 1),
evaluate the strength G of the pairing force for the states 0+ in the nucleus 90 Zr.
b) Applying the pairing interaction with the strength thus evaluated calculate the energies and wave
functions of the states 0+ .
c) Extract from the experimental spectrum of 90 Zr the following interaction matrix elements:
< 1p1/2 0g9/2 ; 4− |V |1p1/2 0g9/2 ; 4− >,
< 1p1/2 0g9/2 ; 5− |V |1p1/2 0g9/2 ; 5− >,
2
< 0g9/2 ; 2+ |V |0g9/2
2
; 2+ >,
< 0g9/2 ; 4 |V |0g9/2 ; 4+ >,
2 + 2

2
< 0g9/2 ; 6+ |V |0g9/2
2
; 6+ > and
< 0g9/2 ; 8 |V |0g9/2 ; 8+ > .
2 + 2

Hint: Extract the single-particle energies in 89 Y from experimental data

http://www.nndc.bnl.gov/nudat2/
You do not need to deduct the influence of the Coulomb field.

Exercise 3:
a) Show that the isospin of two-particle states in a single shell determines whether the space-spin part
of the wave function is symmetric or antisymmetric.
b) Derive Eq. (6.29): Write the state < 12|(nlj)2JM (1/2 1/2)T Tz >a in terms of the corresponding
non-antisymmetrized states. Transform the wave functions back to the neutron-proton scheme for Tz = 0.

Exercise 4:
a) Determine which states are isoscalar and which are isovector in nuclei with two nucleons outside
Z=N=20.
b) Try to find the isoscalar and isovector states in two-hole systems below the core 56
28 Ni28
c) Is it possible to make such identification if the core is Z=20, N=28?

Exercise 5 (not mandatory):

Derive Eqs. (8.32) & (8.33).
Chapter 9
Nuclear pairing

The possibility of pairing in atomic nuclei was first studied by Bohr, Mottelson and Pines and by Belyaev
only one year after the publication of the theory of superconductivity by Bardeen, Cooper and Schrieer
(BCS) [3]. Meanwhile, Bogoliubov developed a microscopic theory of superfluidity and superconductivity
and explored its consequences for nuclear matter. In 1959, Migdal speculated that the interior of neutron
stars might be supefluid.

9.1 Pairing gaps: odd-even binding energy differences

The basic hallmarks of pair condensates are the odd-even staggering in binding energies, the gap in the
excitation spectrum of even systems, and the compressed quasiparticle spectrum in odd systems. To
examine odd-even staggering, it is convenient to define the even and odd neutron pairing gaps as

(3) 1
∆o,Z (N ) = (Eb (Z, N + 1) − 2Eb (Z, N ) + Eb (Z, N − 1)), for N odd, (9.1)
2
(3) 1
∆e,Z (N ) = − (Eb (Z, N + 1) − 2Eb (Z, N ) + Eb (Z, N − 1)), for N even. (9.2)
2
where N and Z are the neutron and proton numbers and Eb is the binding energy of the nucleus. The
proton pairing gaps are defined in a similar way. With the above definition, the gaps are positive for
normal pairing.
One can also see a systematic trend in the gap values as a function of N , namely the gaps get smaller
in heavier nuclei. I will also come back to this behaviour in the theory discussion. Another feature of the
odd-N gap systematics is the occurance of dips at particular values of N . In fact the dips occur adjacent
to the well-known magic numbers N = 28, 50, 82 and 126.
The systematics of the even-N gaps shown in the lower panel is similar with respect to the following:
average values, the fluctuations at each N , and the smooth trend downward with increasing N . However,
the magic number anomolies are now very striking spikes that occur exactly at the magic numbers. Also,
the average values in lighter nuclei appear to be larger for the even-N gaps than for the odd-N . I will
also come back to this feature in the theory section.
The corresponding systematics of proton gaps is shown in Fig. 2. The same qualitative features are
present here as well, but the magic number effects are less pronounced. I do not know of any explanation
of this difference between neutron and proton pairing.
The table below gives some fits to the pairing gap systematics. Shown are the fitted values of the
gap parameterizations and the rms errors of the fits, in units of MeV. The simplest model is a constant

3 3

2.5 Odd N 2.5 Even N

2 2
(MeV)

(MeV)

1.5 1.5

1 1
(3)
(3)

0.5 0.5

0 0

-0.5 -0.5
20 40 60 80 100 120 140 20 40 60 80 100 120 140
Neutron number N Neutron number N

Figure 9.1: Upper panels: odd-N pairing gaps. Lower panels: even-N pairing gaps. Typically, the odd-N
nuclei are less bound than the average of their even-N neighbors by about 1 MeV. However, one sees
that there can be about a factor of two scatter around the average value at a given N .
9 Nuclear pairing 93

3 3

2.5 Odd Z 2.5 Even Z

2 2
(MeV)

e (MeV)
1.5 1.5

1 1

(3)
(3)

0.5 0.5

0 0

-0.5 -0.5
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
Proton number Z Proton number Z

Figure 9.2: Upper panels: odd-Z pairing gaps. Lower panels: even-Z pairing gaps.

gap, ∆(3) = C, shown on the line labeled C. One sees that a typical gap size is 1 MeV, and typical
fluctuations about that are smaller by a factor of 3. Beyond that, there are differences between protons
and neutrons and between the odd and the even gaps. The even gaps are somewhat larger and have
somewhat larger fluctuations, which is to be expected in view of the shell effects exhibited in Fig. 1.
The odd proton gap is smaller than the odd neutron gaps which might be expected from the repulsive
Coulomb contribution to the pairing interaction. There is also a mean-field contribution of the Coulomb
that has opposite signs for even and odd protons. Indeed the even proton gaps are actually larger than
their neutron counterparts.
For the next lines in the table, I come back to the broad trend in Fig. 1, a systematic decrease in
gaps with increasing mass number. It is conventional to describe this with a fractional power dependence,
∆(3) = c/A1/2 . This decreases the rms errors somewhat, but there is no theoretical basis for the fractional
power of A. In the last line I show the result of a two-parameter fit to the functional form ∆(3) = c1 /A+c2 .
This functional form is more justified by theoretical considerations, as will be discussed in the theory
section below.

Basic spectral properties

The other strong signatures of pairing are in excitation spectra. In the simple BCS theory, the lowest
excited states in an even system requires breaking two pairs giving an excitation energy
Eex ≈ 2∆BCS . (9.3)
On the other hand, in the odd particle number system, the quasiparticle level density diverges at the
Fermi energy. This contrasting behavior is very obvious in the nuclear spectrum. As an example, the
isotope chain at proton magic number Z = 50 (the element Sn) has been a favorite for exhibiting and
studying pairing effects. Figure 9.3 shows the low-lying spectra of odd-N members of the chain. One can
see that there are several levels within one MeV of the ground state. The spins and parities of the levels
(not shown) correspond very well with the single-particle orbitals near the neutron Fermi level. However,
the spectrum is very compressed with respect to orbital energies calculated with a shell model potential
well. In contrast, the even members of the chain have no excited states at all within the excitation energy
range displayed in the Figure.
Let us look now look at the global systematics for excitations in the even-even nuclei. The estimate
in Eq. (9.3) is too naive because there can be collective excitations within the gap, as is well-known
from early days of BCS theory [?]. For example, there are longitudinal sound modes in an uncharged
superfluid fermionic liquid. These have a phonon spectrum allowing frequencies within the quasiparticle
gap. One might expect that such modes would be absent in finite systems when the size of the system is
small compared to the coherence length of the pairing field. In fact the situation for nuclear excitations
is much more complicated. However, just for presenting the systematics, we use the right-hand side of
Eq. (9.3) to scale the excitation energies, taking the ratio E2 /2∆, where E2 is the excitation energy and
(3) (3)
∆ is the smaller of ∆eZ and ∆eN .
The scaled excitation energies of the first excited states in even-even are shown in Fig. 9.4. With only
a few exceptions these states have angular momentum and parity quantum numbers J π = 2+ and can
be considered to be collective quadrupolar excitations of the ground state. All of the ratios are smaller
than one, with most in the range 0.1-0.5. The very small excitations in the mass ranges A = 160 − 180
and 220 − 250 correpond to nuclei with static quadrupole deformations.
9 Nuclear pairing 94

1
0.8 Z=50
Energy (MeV)

0.6 Odd N

0.4
0.2
0
50 55 60 65 70 75 80
N

Figure 9.3: Energy levels of odd-N Sn isotopes

0.8

0.6
Ratio

0.4

0.2

0
50 100 150 200 250
Mass number A

Figure 9.4: Energy gap in the excitation spectrum of even-even nuclei, scaled to 2∆(3) . See text for
details.
9 Nuclear pairing 95

The physics underlying these excitations is the softness of a typical nucleus with respect to quadrupolar
deformations. On a qualitative level, the collectivity is similar to the phonon collectivity in the infinite
Fermi gas. A quantitative measure of the collectivity is the sum-rule fraction contained in the excitation,
using the energy-weighted sum rule for some density operator. For the phonon case, the sum rule
fraction approaches 100% when the frequency of the collective mode is small compared to the gap [?].
The collectivity in the nuclear quadrupole excitations is quite different. The sum rule fraction carried
by the lowest 2+ excitation is more or less constant over the entire range of nuclear masses, but it only
about 10% of the total (for isoscalar quadrupole transitions.
Turning to odd-A spectrao, some systematics related to the level density are shown in Fig. ??. The
average excitation energy of the first excited state is plotted for each odd mass number A, averaging over
even values of Z. For comparison, the solid line is the expected spacing in the Fermi gas formula for the
single-particle level density,
dns mkF A
=V ≈ MeV. (9.4)
dE 2π 2 100
The subscript s on Ns indicates that only one spin projection is counted, and kF is the Fermi momentum.
V is the volume of the nucleus, which is (roughly) proportional to the number of nucleons A. One can see
from the Figure that a typical spacing is a factor of 10 smaller than that given by the Fermi gas formula.
Clearly interaction effects are at work to increase the level density near the ground state.
Mean-field theory has made enormous strides in nuclear physics; the self-consistent mean field theory
based on the Hartree-Fock-Bogoliubov approximation and using semi-phenomenological energy function-
als is now the tool of choice for the global description of nuclear structure. It is not my intention to
review this subject since it is well covered elsewhere in this volume.
Neverless, there are number of aspects of nuclear pairing that can be can rather easily understood
using only the more qualitative aspects of pairing theory. Besides the pairing gaps and the effect on level
densities, there are important consequences for two-nucleon transfer reactions and on dynamic properties
such as radioactive decay modes. This section presents an overview of some of these aspects.

Origin of the pairing interaction

It is no surprise that conditions for pairing are satisfied in nuclei. The nuclear interaction between
identical nucleons is strongly attractive in the spin-zero channel, almost to the degree to form a two-
neutron bound state. While this explains the origin at a qualitative level, the many-body aspects of the
nuclear interaction make it difficult to derive a quantitative theory starting from basic interactions. The
progress one has made so far is reviewed in other chapters of this book, so I won’t go into detail here.
But just for perspective, I mention some of the major issues.
I first recall problems with the mean-field interaction to use at the Hartree-Fock level. Most obviously,
the effective interaction between nucleons in the nuclear medium is strongly modified by the Pauli prin-
ciple. The Pauli principle suppresses correlations between nucleons and that in turn make the effective
interaction less attractive. Beyond that, it seems unavoidable to introduce three-body interactions in
a self-consistent mean-field theory. These interactions have two origins. The first is the three-nucleon
interaction arising from sub-nucleon degrees of freedom. It has been convincingly demonstrated that such
interactions are needed to reproduce binding energies of light nuclei and to calculated the bulk properties
of nuclear matter. Besides this more fundamental three-body interaction, there may an induced interac-
tion associated with the short-ranged correlations and their suppression in the many-body environment.
In the popular parameterization of the effective interaction for use in mean-field theory, the three-body
interaction energy has the same order of magitude as the two-body interaction energy. It would thus
seem to be a great oversimplication to ignore the three-body effects in the pairing interaction.
The last issue is the role of the induced interaction associated with low-frequency excitations. We have
seen that the nucleus is rather soft to surface deformations. the virtual excitation of these modes would
contribute to the pairing in exactly the same way that phonon provide an attractive pairing interaction for
the electrons in a superconductor. It may well have the same importance as the two-particle interaction.
Note that if low-frequency phonons were dominant, the energy scale in Eq. (5) would be greatly reduced.
The dynamic properties of an extended fermionic system depend crucially on the presence of a pairing
condensate, changing it from a highly viscous fluid to a superfluid. The effects in nuclei are not quite
as dramatic as in extended systems because the pairing coherence length in nuclei exceeds the size of
the nucleus. Nevertheless, the presence of a highly deformable surface in nuclei requires that pairing be
treated in a dynamical way.
9 Nuclear pairing 96

The most clearly documented dynamic influence of pairing is its effect on the moment of inertia of
deformed nuclei. Without pairing, the rotational spectrum of a deformed fermionic droplet is believed to
follow the spectrum of a rigid rotor,
~2
EJ = J(J + 1). (9.5)
2I
Here ~J is the angular momentum and the moment of inertia I would be close to the rigid value
2 2
I≈ Amhr2 i ≈ A5/3 mr02 . (9.6)
3 5
If the pairing were strong enough to make the coherence length small compared to the size of the system,
the system would be a superfluid having irrotational flow and a corresponding inertial dynamics. What
is somewhat surprising is that the weak pairing that is characteristic of nuclei still has a strong effect on
the inertia.

9.2 The seniority model

For a system with n identical particles in a single j-shell, the binding energy can be solved analytically
in the seniority scheme. The Hamiltonian for such a model can be written as [?],

a†m a†−m a−m0 am0 = −GS+ S−

X
H = −G (9.7)
m,m0 >0

where S represents the quasi-spin operator, and a†m and am are creation and annihilation operators. G
is the strength of the pairing interaction. We have
X (m) X
S+ = s+ = a†m a†−m (9.8)
m m

(m)
X X
S− = (S+ )† = s− = a−m am (9.9)
m m

(m) 1X † 1
(a am + a†−m a−m − 1) = (n − Ω)
X
S0 = s0 = (9.10)
m
2 m m 2

Here, m (m > 0) represents each sub-state level, and a†m and am are creation and annihilation operators.
Ω is the maximal number of pairs in the single j-shell (Ω = j + 21 ). The eigenvalues S of total quasi-spin
are: S = 21 |N − Ω|, · · · , 12 Ω − 1, 12 Ω, and the energy eigenvalues of H are given by:
 
1 2 1
E(S) = −G S · (S + 1) − (N − Ω) + (N − Ω) (9.11)
4 2

Or we can introduce seniority quantum number s given by s = Ω − 2S, the energy of the state with
seniority v can be written as
n−v
E(v) = −G (2j + 3 − n − v) (9.12)
4
n(n − 1) v(v − 1) 1
= G− G − (n − v)(j + 1)G
4 4 2
If we assume v = 0 for the ground state of even-even system and v = 1 for that of the odd system. The
expression above can be simplified as,

n(n − 1) hni
E(v) = G− (j + 1)G, (9.13)
4 2
where [n/2] denotes the largest integer not exceeding n/2 and corresponds to the total number of v = 0
pairs.
9 Nuclear pairing 97

9.3 Basics of pairing correlations

from arXiv:1206.2600
In quantum mechanics of finite many-fermion systems, pairing correlations are best described in terms
of number operators N̂µ = a+ µ aµ , where µ represents any suitable set of single-particle (s.p.) quantum
numbers. Thus we may have, e.g., µ ≡ kσ for plane waves of spin- 12 particles (electrons); µ ≡ rστ for
spin- 21 and isospin- 12 nucleons localized in space at position r; and µ ≡ n, `, j, m for fermions moving
in a spherical potential well. Since N̂µ2 = N̂µ , the number operators are projective; hence, one can –
at least in principle – devise an experiment that would project any quantum many-fermion state |Ψi
into its component with exactly one fermion occupying state µ. As the rules of quantum mechanics
stipulate, any such individual measurement can only give 0 or 1 (these are the eigenvalues of N̂µ ),
whereas performing such measurements many times, one could experimentally determine the occupation
probabilities vµ2 = hΨ|N̂µ |Ψi.
Along such lines, we can devise an experiment that would determine the simultaneous presence of
two fermions in different orthogonal s.p. states µ and ν. Since the corresponding number operators N̂µ
and N̂ν commute, one can legitimately ask quantum mechanical questions about one-particle occupation
probabilities vµ2 and vν2 , as well as about the two-particle occupation. The latter one reflects the simul-
2
taneous presence of two fermions in state |Ψi: vµν = hΨ|N̂µ N̂ν |Ψi. In this way, one can experimentally
determine the pairing correlation between states µ and ν as the excess probability
2
Pµν = vµν − vµ2 vν2 , (9.14)

of finding two fermions simultaneously over that of finding them in independent, or sequential, measure-
ments. Such a definition of pairing is independent of its coherence, collectivity, nature of quasiparticles,
symmetry breaking, thermodynamic limit, or many other notions that are often associated with the phe-
nomenon of pairing. In terms of occupations, pairing can be viewed as a measurable property of any
quantum many-fermion state.
Obviously, no pairing correlations are present in a quantum state that is an eigenstate of N̂µ or N̂ν ,
such as the Slater determinant. The beauty of the BCS ansatz is in providing us with a model N -fermion
state, in which pairing correlations are explicitly incorporated:
!N/2
X
|ΦN i = N N sµ zµ a+ +
µ̃ aµ |0i, (9.15)
µ>0

where the summation µ > 0 runs over the representatives of pairs (µ̃, µ) of s.p. states (that is, any one
state of the pair is included in the sum, but not both), zµ̃ = zµ are real positive numbers, sµ̃ = −sµ are
arbitrary complex phase factors, and N N is the overall normalization factor.
It now becomes a matter of technical convenience to employ a particle-number mixed state,
∞
!
X |ΦN i X
+ +
|Φi = N = N exp sµ zµ aµ̃ aµ |0i, (9.16)
N N (N/2)! µ>0
N =0,2,4,...

in which the pairing correlations (9.14) are:

zµ2 1
Pµν = vµ2 u2ν δµ̃ν for vµ2 = and u2ν = . (9.17)
1 + zµ2 1 + zν2

In terms of the s.p. occupations, the state |Φi assumes the standard BCS form:
Y 
|Φi = uµ + sµ vµ a+ +
µ̃ µ |0i.
a (9.18)
µ>0

In this many-fermion state, the s.p. states µ̃ and µ are paired, that is, |Φi can be viewed as a pair-
condensate. For zµ =1, the pairing correlation Pµ̃µ (9.14) equals 1/4; in fact, in this state, it is twice more
2
likely to find a pair of fermions (vµ̃µ =1/2) than to find these two fermions independently (vµ̃2 vµ2 =1/4).
For particle-number conserving states (9.15), the occupation numbers can be calculated numerically;
qualitatively the results are fairly similar, especially for large numbers of particles.
9 Nuclear pairing 98

At this point, we note that the most general pair-condensate state (9.16) has the form of the Thouless
state, !
X
∗ + +
|Φi = N exp 12 Zνµ aν aµ |0i, (9.19)
νµ

in which pairs (µ̃, µ) do not appear explicitly. However, there always exists a unitary transformation
U0 of the antisymmetric matrix Z that brings it to the canonical form (U0+ Z ∗ U0∗ )νµ = sµ zµ δµ̃ν (the
Bloch-Messiah-Zumino theorem[?, ?]). Therefore, pairs are present in any arbitrary Thouless state (the
so-called canonical pairs), and they can be made explicitly visible by a simple basis transformation.
The canonical pairs exist independently of any symmetry of the Thouless state. In the particular case
of a time-reversal-symmetric state, T̂ |Φi = |Φi, they can be associated with the time-reversed s.p. states,
µ̃ ≡ µ̄. The ground-states of even-even nuclei can be described in this manner. However, the appearance
of pairing phase does not hinge on this particular symmetry – states in rotating nuclei (in which time-
reversal symmetry is manifestly broken) can also be paired. In this latter case the canonical states are
less useful, because they cannot be directly associated with the eigenstates of the HFB Hamiltonian.

Hartree-Fock-Bogoliubov theory
The simplest route to the HFB theory is to employ the variational principle to a two-body Hamiltonian
using Thouless states (9.19) as trial wave functions. The variation of the average energy with respect
to the antisymmetric matrix Z results in the HFB equation in the matrix representation, HU = UE, or
explicitly,
U V∗ U V∗
     
T +Γ ∆ E 0
= , (9.20)
−∆∗ −T ∗ − Γ∗ V U∗ V U∗ 0 −E
where Tµν is the matrix of the one-body kinetic energy, Γµν = µ0 ν 0 Vµµ0 ;νν 0 ρν 0 µ0 and ∆µµ0 = 21 νν 0 Vµµ0 ;νν 0 κνν 0
P P
are the so-called particle-hole and particle-particle mean fields, respectively, obtained by averaging two-
body matrix elements Vµµ0 ;νν 0 with respect to the density matrix ρν 0 µ0 = hΦ|a+ µ0 aν 0 |Φi and pairing tensor
κνν 0 = hΦ|aν 0 aν |Φi, and E is the diagonal matrix of quasiparticle energies[?].
The matrices H and U are referred to as the HFB Hamiltonian and Bogoliubov transformation,
respectively, and columns of U (eigenstates of H) are vectors of quasiparticle states. The HFB equation
(9.20) possesses the quasiparticle-quasihole symmetry. Namely, for each quasiparticle state χα (the α-th
column of U) and energy Eα there exists a quasihole state φα of opposite energy −Eα ,
   ∗ 
Uµα Vµα
χα = , φα = ∗ . (9.21)
Vµα Uµα

That is, the spectrum of H is composed of pairs of states with opposite energies. In most cases, the lowest
total energy is obtained by using the eigenstates with Eα > 0 as quasiparticles χα and those with Eα < 0
as quasiholes φα , that is, by occupying the negative-energy eigenstates. States χα and φα can usually be
related through a self-consistent discrete symmetry, such as time reversal, signature, or simplex.[?, ?, ?].
The HFB equation (9.20) is also valid in a more general case, when the total energy is not equal to
the average of any many-body Hamiltonian. Within the DFT, it stems from the minimization of the
binding energy given by an EDF E(ρ, κ, κ∗ ), subject to the condition of the generalized density matrix
being projective, that is, R2 = R for
  ∗ T
V ∗U T
   ∗ 
V T UT

X
ρ κ V V V
R= = = = φα φ +
α. (9.22)
−κ∗ 1 − ρ∗ U ∗V T U ∗U T U∗
α

In this case, the mean fields are obtained as functional derivatives of EDF: Γµν = ∂E/∂ρνµ and ∆µµ0 =
∂E/∂κ∗µµ0 . As is the case in DFT, densities (here the density matrix and pairing tensor) become the
fundamental degrees of freedom, whereas the state |Φi acquires the meaning of an auxiliary entity (the
Kohn-Sham state[?]). Indeed, for any arbitrary generalized density matrix R (9.22), one can always
find the corresponding state |Φi.  For that, one determines the Bogoliubov transformation U as the
0 0
matrix of its eigenvectors, RU = U ; the Thouless state |Φi (9.19) corresponds to Z = V U −1 .
0 1
Consequently, the paired state |Φi of DFT is not interpreted as a wave function of the system – it only
serves as a model for determining one-body densities. Nonetheless, these densities are interpreted as
those associated with the (unknown) exact eigenstate of the system.
9 Nuclear pairing 99

Unrestricted variations of the EDF are not meaningful. Indeed, since Thouless states (9.16) are
mixtures of components with different particle numbers, absolute minima will usually correspond to
average particle numbers that are unrelated to those one would like to describe. In particular, for self-
bound systems governed by attractive two-body forces (nuclei), by adding more and more particles one
could infinitely decrease the total energy of the system. Therefore, only constrained variations make sense,
that is, one has to minimize not the total energy E(ρ, κ, κ∗ ), but the so-called Routhian, E 0 (ρ, κ, κ∗ ) =
E(ρ, κ, κ∗ ) + C(ρ), where C is a suitably chosen penalty functional, ensuring that the minimum appears
at prescribed average values of one-body operators. In particular, the average total number of particles
can be constrained by C(ρ) = −λhΦ|N̂ |Φi = −λTr(ρ) (linear constraint) or C(ρ) = CN [Tr(ρ) − N0 ]2
(quadratic constraint),[?, ?] where λ becomes the Fermi energy corresponding to N0 fermions.
For different systems and for different applications, various constraints C(ρ) can be implemented; for
example, in nuclei one can simultaneously constrain numbers of protons and neutrons, as well as multipole
moments of matter or charge distributions. When the total energy is a concave function of relevant one-
body average values, quadratic constraints are mandatory[?, ?]. The minimization of E 0 (ρ, κ, κ∗ ) requires
0
solving the HFB equation for the quasiparticle Routhian
 H , which, for the simplest case of the constraint
1 0
on the total particle number, reads H0 = H − λ .
0 −1

9.4 Exact solutions for pairing interactions

Taken from arXiv:1204.2950
The first breakthrough towards a microscopic description of the superconducting phenomenon was
due to Cooper[1], who in 1956 showed that a single pair of electrons on top of an inert Fermi sea could be
bound by an infinitesimal attractive interaction. The search for a many-body wave function describing a
fraction of correlated and overlapping pairs mixed with a Fermi sea was a key goal for the rest of that year.
Schrieffer came up with a solution at the beginning of 1957 and the BCS team (Bardeen, Cooper and
Schrieffer) started an intensive and fruitful collaboration to explain quantitatively many superconducting
properties from the associated BCS wave function. This led to the famous BCS paper [2] which provided
a complete microscopic explanation of superconductivity.
The success of the BCS theory quickly spread to other quantum many-body systems, including the
atomic nucleus. In the summer of 1957, David Pines visited the Niels Bohr Institute and gave a series of
seminars about the yet unpublished BCS theory. Soon thereafter Bohr, Mottelson and Pines published
a paper [3] suggesting that the gaps observed in even-even nuclei could be due to superconducting
correlations. They noted, however, that these effects should be strongly influenced by the finite size of
the nucleus. Since then, and up to the present, number projection and in general symmetry restoration in
the BCS and Hartree-Fock-Bogoliubov approximations have been important issues in nuclear structure.
At the beginning of the sixties, while several groups were developing numerical techniques for number-
projected BCS calculations [4, 5], Richardson provided an exact solution for the reduced BCS Hamiltonian
[6, 7]. In spite of the importance of his exact solution, this work did not have much impact in nuclear
physics with just a few exceptions. Later on, his exact solution was rediscovered in the framework
of ultrasmall superconducting grains [8] where BCS and number-projected BCS were unable to describe
appropriately the crossover from superconductivity to a normal metal as a function of the grain size. Since
then, there has been a flurry of work extending the Richardson exact solution to families of exactly-solvable
models, now called the Richardson-Gaudin (RG) models [9], and applying these models to different areas
of quantum many-body physics including mesoscopic systems, condensed matter, quantum optics, cold
atomic gases, quantum dots and nuclear structure [10]. In this paper, we review Richardson’s solution, its
generalization to the exactly-solvable RG models and discuss the applications of these models in nuclear
physics.

The Richardson solution of the reduced BCS hamiltonian

We will focus on a pairing Hamiltonian with constant strength G acting in a space of doubly-degenerate
time-reversed states (k, k̄), X † X † †
HP = k ck ck − G ck ck ck0 ck0 , (9.23)
k k,k0

where k are the single-particle energies for the doubly-degenerate orbits k, k̄.
9 Nuclear pairing 100

Cooper considered the addition of a pair of fermions with an attractive pairing interaction on top of
an inert Fermi sea (FS) under the influence of this Hamiltonian. He showed that the pair eigenstate is
1
c† c† |F Si ,
X
|ΨCooper i = (9.24)
2k − E k k
k>kF

where E is the energy eigenvalue. Cooper found that for any attractive value of G, the Fermi sea is
unstable against the formation of such bound pairs. Therefore, an approach that takes into account a
fraction of these correlated pairs mixed with a Fermi sea should be able to describe the superconducting
phenomenon.
The BCS approach followed a somewhat different path to the one suggested by Cooper, defining
instead a variational wave function as a coherent state of pairs that are averaged over the whole system,
†
|ΨBCS i = eΓ |0i , (9.25)
† †
where Γ† =
P
k zk ck ckis the coherent pair. Though errors due to the non-conservation of particle number
in (9.25) are negligible when the number of pairs is sufficiently large, they can be important in such finite
systems as atomic nuclei [3]. To accommodate these effects, number-projected BCS (PBCS) [4] considers
a condensate of pairs of the form

M
|ΨP BCS i = Γ† |0i , (9.26)
†
where M is the number of pairs and Γ has the same form as in BCS.
Richardson [6] proposed an ansatz for the exact solution of the pairing Hamiltonian (9.23) that followed
closely Cooper’s original idea. For a system with 2M + ν particles, with ν of these particles unpaired,
his ansatz involves a state of the form

|Ψi = B1† B2† · · · BM

†
|νi , (9.27)

where the collective pair operators Bᆠhave the form found by Cooper for the one-pair problem,
L
1
c† c† .
X
Bᆠ= (9.28)
2εk − Eα k k
k=1

Here L is the number of single-particle levels and

|νi ≡ |ν1 , ν2 · · · , νL i (9.29)

P
is a state of ν unpaired fermions (ν = k νk , with νk = 1 or 0) defined by ck ck |νi = 0, and nk |νi = νk |νi.
In the one-pair problem, the quantities Eα that enter (9.28) are the eigenvalues of the pairing Hamil-
tonian, i.e., the pair energies. Richardson proposed to use the M pair energies Eα in the many-body
wave function of Eqs. (9.27, 9.28) as parameters which are chosen to fulfill the eigenvalue equation
HP |Ψi = E |Ψi. He showed that this is the case if the pair energies satisfy a set of M non-linear coupled
equations
L M
X 1 − νk X 1
1−G − 2G =0, (9.30)
2εk − Eα Eβ − Eα
k=1 β(6=α)=1

which are now called the Richardson equations. The second term represents the interaction between
particles in a given pair and the third term represents the interaction between pairs. The associated
eigenvalues of H are given by
XL M
X
E= εk νk + Eα , (9.31)
k=1 α=1

namely as a sum of the pair energies.

Each independent solution of the set of Richardson equations defines a set of M pair energies that
completely characterizes a particular eigenstate (9.27, 9.28). The complete set of eigenstates of the pairing
Hamiltonian can be obtained in this way. The ground state solution is the energetically lowest solution in
the ν = 0 or ν = 1 sector, depending on whether the system has an even or an odd number of particles,
respectively.
9 Nuclear pairing 101

There are a couple of points that should be noted here. First, in contrast to the BCS solution, each
Cooper pair Bα† is distinct. Second, if one of the pair energies Eα is complex, then its complex-conjugate
Eα∗ is also a solution. From this latter point we see that |Ψi preserves time-reversal invariance.
On inspection of the Richardson pair (9.28), we see that a pair energy that is close to a particular 2k ,
i.e. close to the energy of an unperturbed pair, is dominated by this particular configuration and thus
defines an uncorrelated pair. In contrast, a pair energy that lies sufficiently far away in the complex plane
produces a correlated Cooper pair. This is to be contrasted with the single BCS coherent pair, which has
amplitude zk = vk /uk and which mixes correlated and uncorrelated pairs over the whole system.

Generalization to the Richardson-Gaudin class of integrable models

In this section, we discuss how to generalize the standard pairing model, which as we have seen is exactly
solvable, to a wider variety of exactly-solvable models, the so-called Richardson-Gaudin models[11], all
of which are based on the SU (2) algebra. We first introduce the generators of SU (2), using a basis more
familiar to nuclear structure,
!
1 X † X † †
0
Kj = ajm ajm − Ωj , Kj+ = ajm ajm , Kj− = (Kj+ )† . (9.32)
2 m m

Here a†jm creates a fermion in single-particle state jm, jm denotes the time reverse of jm, and Ωj =
j + 12 is the pair degeneracy of orbit j. These operators fulfill the SU (2) algebra [Kj+ , Kj−0 ] = 2δjj 0 Kj0 ,
[Kj0 , Kj±0 ] = ±δjj 0 Kj± .
We now consider a general set of L Hermitian and number-conserving operators that can be built up
from the generators of SU (2) with linear and quadratic terms,
X  Xij 
0 + − − + 0 0


Ri = Ki + 2g Ki Kj + Ki Kj + Yij Ki Kj . (9.33)
2
j(6=i)

Following Gaudin[12], we then look for the conditions that the matrices X and Y must satisfy in order
that the R operators commute with one another. It turns out that there are essentially two families of
solutions, referred to as the rational and hyperbolic families, respectively.
i. The rational family
1
Xij = Yij = (9.34)
ηi − ηj
ii. The hyberbolic family
√
ηi ηj ηi + ηj
Xij = 2 , Yij = (9.35)
ηi − ηj ηi − ηj
Here the set of L parameters ηi are free real numbers.
The traditional pairing model is an example P of the rational family. It can be obtained as a linear
combination of the integrals of motion, HP = j εj Rj (εj ), with ηj = εj .
The complete set of eigenstates of the rational integrals of motion is given by the Richardson ansatz
(9.27, 9.28). This fact led Gaudin [12] to try to relate his integrable models to the BCS Hamiltonian
without success. The proof of integrability of the BCS Hamiltonian was found later in ref. [13]. We will
not present the general solution of the two integrable families here, referring the reader to refs. [11, 9, 10].
The key point is that any Hamiltonian that can be expressed as a linear combination of the R operators
can be treated exactly using this method. In the following sections, we discuss nuclear applications of
the standard pairing model and of a new model based on the hyberbolic family.

Applications of the Richardson solution to pairing in nuclear physics

Richardson himself started to explore analytically the exact solution in nuclear structure for few pairs
outside a doubly-magic core[14, 15]. He also proposed a numerical method to solve the equations for
systems with equidistant levels[16], a model that was subsequently used as a benchmark to test many-
body approximations [17]. However, the first application of the Richardson solution to a real nuclear
system was reported by Andersson and Krumlinde [18] in 1977. They studied the properties of high-
spin states in 152 Dy using an oblate deformed oscillator potential and including the effects of pairing
9 Nuclear pairing 102

at several different levels of approximation. They compared the results when pairing was treated with
the traditional BCS approximation, when it was treated in PBCS approximation (using the saddle point
approximation) and when it was treated exactly using the Richardson method.
Following that early work, there were sporadic references to the Richardson method but no realistic
studies of atomic nuclei until just a few years ago. In 2007, Dussel et al. [19] reported a systematic study
of pairing correlations in the even Sm isotopes, from 144 Sm through 158 Sm, using the self-consistent
deformed Hartree Fock+BCS method. The calculations made use of the density-dependent Skyrme
force, SLy4, and treated pairing correlations using a pairing force with constant strength G assuming
axial symmetry and taking into account 11 major shells.
Using the results at self-consistency to define the HF mean field, pairing effects within that mean field
were then considered using the alternative number-conserving PBCS approach and the exact Richardson
approach. In this way it was possible to directly compare the three approaches to pairing with the
same pairing Hamiltonian, a primary focus of the study. It should be noted here that the Hilbert space
dimensions associated with the residual neutron pairing Hamiltonian is of the order of 3.9 × 1053 for
154
Sm, whereas the exact Richardson approach requires the solution of a coupled set of 46 non-linear
equations.
In the one semi-magic nucleus 144 Sm that was studied, the principal correlation effects arise when
projection is included, taking the system from one that is normal at the level of BCS to one with
substantial pairing correlations. Treating pairing exactly provides a further modest increase in pairing
correlations of about 0.3 M eV . In non-semi-magic nuclei, the effect on the pairing correlation energy
of the exact solution is significantly more pronounced. While there too number projection provides a
substantial lowering of the energy, it now misses about 1 M eV of the exact correlation energy that derives
from the Richardson solution.
The exact Richardson solution was also used to study the gradual emergence of superconductivity
in the Sn isotopes [22]. By making use of an exact mapping between the Richardson equations and
a classical electrostatic problem in two dimensions, it was possible to get a physical picture of how
superconductivity develops as a function of the pairing strength. In particular, as the pairing strength is
increased the pair energies gradually merge into larger structures in the complex plane as pair correlations
gradually overcome single-particle effects.
More recently, the Richardson solution has been applied to the treatment of pair correlations involving
the continuum. The first work by Hasegawa and Kaneko [23] considered only the effect of resonances
in the continuum and as a result obtained complex energies even for the bound states of the system.
Subsequent work by Id Betan [24, 25] included the effects of the true continuum. The most recent paper
[25] treated nuclear chains that include both bound and unbound systems, e.g. the even-A Carbon
isotopes up to 28 C. When the system is bound, the pair energies that contribute to the ground state
occur in complex conjugate pairs, thus preserving the real nature of the ground state energy. Once the
system becomes unbound this ceases to be the case. Now the pair energies that contribute to the ground
state do not occur in complex conjugate pairs, explaining how a width arises in the energy of an unbound
system within the Richardson approach.

The hyperbolic model

The hyperbolic family of models did not find a physical realization until very recently when it was shown
that they could model a p-wave pairing Hamiltonian in a 2-dimensional lattice [26], such that it was
possible to study with the exact solution an exotic phase diagram having a non-trivial topological phase
and a third-order quantum phase transition [27]. Immediately thereafter, it was shown that the hyperbolic
family gives rise to a separable pairing Hamiltonian with 2 free parameters that can be adjusted to
reproduce the properties of heavy nuclei as described by a Gogny HFB treatment [28]. Both applications
are based on a simple linear combination of hyperbolic integrals which give rise to the separable pairing
Hamiltonian X X√
H= ηi Ki0 − G ηi ηi0 Ki+ Ki−0 . (9.36)
i i,i0

If we interpret the parameters ηi as single-particle energies corresponding to a nuclear mean-field

potential, the pairing interaction has the unphysical behavior of increasing in strength with energy. In
order to reverse this unwanted effect, we define ηi = 2(εi − α), where the free parameter α plays the role
of an energy cutoff and εi is the single-particle energy of the mean-field level i. Making use of the pair
9 Nuclear pairing 103

representation of SU (2), the exactly-solvable pairing Hamiltonian (9.36) takes the form
X  † 
εi ci ci + ci† ci − 2G (α − εi ) (α − εi0 ) c†i ci† ci0 ci0 ,
Xp
H= (9.37)
i ii0
P P
with eigenvalues E = 2αM + i εi νi + β Eβ . The pair energies Eβ correspond to a solution of the set
of non-linear Richardson equations
1X 1 X 1 Q
− = , (9.38)
2 i ηi − Eβ 0
Eβ 0 − Eβ Eβ
β (6=β)

1
where Q = 2G − L2 + M − 1. Each particular solution of Eq. (9.38) defines a unique eigenstate.
Due to the separable character of the hyperbolic Hamiltonian, in BCS approximation
√
the gaps ∆i =
√ P √ √
2G α − εi i0 α − εi0 ui0 vi0 = ∆ α − εi and the pairing tensor ui vi = √ ∆ 2α−εi 2
have a very
2 (εi −µ) +(α−εi )∆
restricted form. In order to test the validity of the exactly solvable Hamiltonian (9.37) we take the single-
particle energies εi from the HF energies of a Gogny HFB calculation and we fit the parameters α and
G to the gaps and pairing tensor in the HF basis. Figure ?? shows the comparison for protons in 238 U
between the Gogny HFB results in the HF basis and the BCS approximation of the hyperbolic model.
From these results we extracted the values α = 25.25 M eV and G = 2 × 10−3 M eV . The valence space
determined by the cutoff α corresponds to 148 levels with 46 proton pairs. The size of the Hamiltonian
in this space is 4.83 × 1038 , well beyond the limits of exact diagonalization. However, the integrability
of the hyperbolic model provides an exact solution by solving a set of 46 non-linear coupled equations.
Moreover, the exact solution shows a gain in correlation of more than 2 M eV suggesting the importance
of taking into account correlations beyond mean-field.
Up to now, we have restricted our discussion to RG models that are based on the compact rank-1
SU (2) pair algebra. The method of constructing RG models can be extended to the non-compact rank-1
SU (1, 1) algebra as well, whereby pairing in bosonic systems[29] is described in complete analogy with
the SU (2) case. An early application to the SO(6) to U (5) line of integrability of the Interacting Boson
Model (IBM) was reported in ref. [30], with the exact solution being obtained there directly using an
infinite dimensional algebraic technique. Further work on the IBM using the integrable SU (1, 1) RG
model[31] including high-spin bosons (d, g, · · · ) revealed a particular feature of the repulsive boson
pairing interaction that seems to provide a new mechanism for the enhancement of s − d dominance,
giving further support for the validity of the s − d Interacting Boson Model.
The RG models are not constrained to rank-1 algebras. They can be extended to any semi-simple Lie
algebra [32]. Richardson himself studied some restricted solutions of the T=1 pairing model [33] and the
T=0,1 pairing model [34]. As a general statement, the reduced pairing Hamiltonian is exactly solvable
for any multi-component system. The first step in finding an exact solution is to identify the Lie algebra
of the commuting pair operators and then to specialize the general solution given in [32]. One has to
keep in mind that while the SU (2) RG model has a single set of unknown parameters, the pair energies,
larger rank algebras have as many sets of unknown parameters as the rank of the algebra. Therefore, the
higher the rank of the algebra, the greater is the complexity of the solution. Several pairing Hamiltonians
with relevance to nuclear physics have been studied in the last few years.
i. The rank-2 SO(5) RG model[35] describes T=1 proton-neutron pairing with non-degenerate single
particle levels. The exact solution has two sets of spectral parameters, the pair energies and a second
set associated with the SU (2) isospin subalgebra. In spite of the greater complexity, it was possible to
solve exactly a T=1 pairing Hamiltonian for the nucleus 64 Ge using a 40 Ca core, with a Hilbert space
dimension well beyond the limits of exact diagonalization.
ii. The rank-3 SO(6) RG model [36] describes color pairing, i. e. pairing between three-component
fermions. The exact Richardson equations have three sets of spectral parameters, of which one correspond
the the pair energies and the other two are responsible for the different couplings within the SU (3) color
subalgebra. The model has been used to study the phase diagram of polarized three-component fermion
atomic gases. However, it could in principle be exploited to describe non-relativistic quark systems.
iii. With increasing complexity, the rank-4 SO(8) RG model [37] describes either T=0,1 proton-
neutron pairing or four-component fermion gases. It contains four sets of spectral parameters. The
model has been used to study alpha-like structures represented by clusters in the parameter space, and
how these clusters dissolve into like-particle pairs with increasing isospin.
iv. The rank-2 non-compact SO(3, 2) algebra generalizes the bosonic RG models to systems of inter-
acting proton and neutron bosons[38]. The model describes the IBM2 in the line of integrability between
9 Nuclear pairing 104

vibrational and γ-soft nuclei. The exact solution has been employed to study the influence of high-spin
f and g bosons in the low-energy spectrum
A key feature of the Richardson-Gaudin integrable models, is that they transform the diagonalization
of the hamiltonian matrix, whose dimension grows exponentially with the size of the system, to the
solution of a set of M coupled non-linear equations where M is the number of pairs. This makes it
possible to treat problems that could otherwise not be treated and in doing so to obtain information that
is otherwise inaccessible. For example, we reported an application of the rational RG pairing model to the
even-mass Sm isotopes, where the size of the Hilbert space would exceed 1053 states, and an application
of the hyperbolic RG pairing model to U 238 , where the size of the Hilbert space would exceed 1038 states.
In both cases, substantial gains in correlation energy were found when the problem was treated exactly.
The exactly solvable RG Hamiltonians also provide excellent benchmarks for testing approximations
beyond HFB in realistic situations both for even-even and odd-mass nuclei. Moreover, a self-consistent
HF plus exact pairing approach could in principle be implemented to describe large regions of the table of
nuclides. It might be possible to extend such a self-consistent approach to the O(5) RG model, providing
in this way a better description of those nuclei with N ∼ Z in which T=1 proton-neutron pairing
correlations are expected to play a significant role. Unfortunately, the SO(8) T=0,1 RG model cannot
accommodate the spin-orbit splitting in the single-particle energies. Nevertheless, this model could play
an important role in helping to understand quartet clusterization and quartet condensation in nuclear
and cold atom systems. Finally, extension of the RG models to include the effects of the continuum seems
to be an especially promising avenue to explore the physics of weakly-bound nuclei.

9.5 Homework problems

Exercise 1: Evaluate the experimental pairing gaps in Ca isotopes based on Eq. (9.1) and compare
them with the predictions of the seniority model, Eq. (9.13).
 Bibliography

[1] L. N. Cooper, Bound Electron Pairs in a Degenerate Fermi Gas, Phys. Rev. 104, 1189–1190, (1956).
[2] J. Bardeen, L. N. Cooper, and J. R. Schrieffer, Theory of Superconductivity, Phys. Rev. 108, 1175–
1204, (1957).

[3] A. Bohr, B. R. Mottelson, and D. Pines, Possible Analogy between the Excitation Spectra of Nuclei
and those of the Superconducting Metallic State, Phys. Rev. 110, 936–938, (1958).
[4] A. K. Kerman, R. D. Lawson, and M. H. Macfarlane, Accuracy of the Superconductivity Approxi-
mation for Pairing Forces in Nuclei, Phys. Rev. 124, 162–167, (1961).
[5] K. Dietrich, H. J. Mang, and J. H. Pradal, Conservation of Particle Number in the Nuclear Pairing
Model, Phys. Rev. 135, 22–34, (1964).
[6] R. W. Richardson, A Restricted Class of Exact Eigenstates of the Pairing-Force Hamiltonian, Phys.
Lett. 3, 277–279, (1963).
[7] R. W. Richardson, Exact Eigenstates of Pairing-Force Hamiltonian, Nucl. Phys. 52, 221–238, (1964).

[8] G. Sierra, J. Dukelsky, G. G. Dussel, J. von Delft, and F. Braun, Exact study of the effect of level
statistics in ultrasmall superconducting grains, Phys. Rev. B 61, 11890–11893, (2000).
[9] J. Dukelsky, S. Pittel, and G. Sierra, Exactly solvable Richardson-Gaudin models for many-body
quantum systems, Rev. Mod. Phys. 76, 643–662 (2004).

[10] G. Ortiz, R. Somma, J. Dukelsky, and S. Rombouts, Exactly-solvable models derived from a gener-
alized Gaudin algebra, Nucl. Phys. B 707, 421–457 (2005).
[11] J. Dukelsky, C. Esebbag, and P. Schuck, Class of Exactly Solvable Pairing Models, Phys. Rev. Lett
87, 066403 1–4 (2001).
[12] M. Gaudin, Diagonalization of a Class of Spin Hamiltonian, J. Phys. (Paris) 37, 1087–1098 (1976).

[13] M. C. Cambiaggio,A. M. F. Rivas, and M. Saraceno, Integrability of the pairing Hamiltonian, Nuc.
Phys. A 624, 157–167 (1997).
[14] R. W. Richardson, Application to the Exact Theory of the Pairing Model to some Even Isotopes of
Lead, Phys. Lett. 5, 82–84, (1964).
206 204 202
[15] R. W. Richardson and N. Sherman, Pairing Models of P b, P b and P b, Nuc. Phys. 52,
253–268, (1964).
[16] R. W. Richardson, Numerical Study of 8-32 Particle Eigenstates of Pairing Hamiltonian, Phys. Rev.
141, 949–956, (1966).

[17] J. Bang and J. Krumlinde, Model Calculations with Pairing Forces, Nucl. Phys. A 141, 18–32,
(1970).
[18] C. G. Andersson and J. Krumlinde, Oblate High-Spin Isomers, Nucl. Phys. A 291, 21–44, (1977).
[19] G. G. Dussel, S. Pittel, J. Dukelsky, P. Sarriguren, Cooper pairs in atomic nuclei, Phys. Rev. C 76
011302 1–5, (2007).

[20] Similar results has been obtained for a 3D homogeneous diluted Fermi gas in the BCS phase. G.
Ortiz and J. Dukelsky, BCS-to-BEC crossover from the exact BCS solution, Phys. Rev. A 72 043611
1–5, (2005).
[21] M. Matsuo, Spatial structure of neutron Cooper pair in low density uniform matter, Phys. Rev. C
73 044309 1–16, (2005); and Matsuo’s contribution to this Volume.
9 Bibliography 106

[22] J. Dukelsky, C. Esebbag, S. Pittel, Electrostatic mapping of nuclear pairing, Phys. Rev. Lett. 88
062501 1–4, (2002).
[23] M. Hasgawa and K. Kaneko, Effects of resonant single-particle states on pairing correlations, Phys.
Rev. C 67, 024304 1–4, (2003).
[24] R. Id Betan, Using continuum level density in the pairing Hamiltonian: BCS and exact solutions,
Nucl. Phys. A 879 14–24, (2012).
[25] R. Id Betan, Exact eigenvalues of the pairing Hamiltonian using continuum level density, Nucl-th
arxiv:1202.3986 (2012).
[26] M. Ibañez, J. Links, G. Sierra, and S-Y Zhao, Exactly solvable pairing model for superconductors
with px+ipy-wave symmetry, Phys. Rev. B 79 180501 1–4, (2009).
[27] S. M. A. Rombouts, J. Dukelsky, and G. Ortiz, Quantum phase diagram of the integrable px+ipy
fermionic superfluid, Phys. Rev. B 82 224510 1–4, (2010).

[28] J. Dukelsky, S. Lerma H., L. M. Robledo, R. Rodriguez-Guzman, and S. M. A. Rombouts, Exactly

solvable Hamiltonian for heavy nuclei, Phys. Rev. C 84 061301 1–4, (2011).
[29] J. Dukelsky and P. Schuck, Condensate Fragmentation in a New Exactly Solvable Model for Confined
Bosons, Phys. Rev. Lett. 86, 4207–4210 (2001).

[30] Feng Pan and J.P. Draayer, New algebraic solutions for SO(6) to U(5) transitional nuclei in the
Interacting Boson Model, Nucl. Phys. A 636, 156-168 (1998).
[31] J. Dukelsky and S. Pittel, New Mechanism for the Enhancement of sd Dominance in Interacting
Boson Models, Phys. Rev. Lett. 86, 4791–4794 (2001).

[32] M. Asorey, F. Falceto, and G. Sierra, ChernSimons theory and BCS superconductivity, Nucl. Phys.
B 622, 593–614, (2002).
[33] R. W. Richardson, Eigenstates of the J=0 T=1 Charge-Independent Pairing Hamiltonian, Phys.
Rev. 144, 874–883, (1966).
[34] R. W. Richardson, Eigenstates of the L=0 T=1 Charge- and Spin-Independent Pairing Hamiltonian,
Phys. Rev. 159, 792–805, (1967).
[35] J. Dukelsky, V. G. Gueorguiev, P. Van Isacker, S. Dimitrova, B. Errea, and S. Lerma H., Exact
Solution of the Isovector Neutron-Proton Pairing Hamiltonian, Phys. Rev. Lett. 76, 072503 1–4,
(2006).

[36] B. Errea, J. Dukelsky, and G. Ortiz, Breached pairing in trapped three-color atomic Fermi gases,
Phys. Rev. A. 79, 051603 1–4, (2009).
[37] S. Lerma H., B. Errea, J. Dukelsky, and W. Satula, Exact Solution of the Spin-Isospin Proton-
Neutron Pairing Hamiltonian, Phys. Rev. Lett. 79, 032501 1–4 (2007).
[38] S. Lerma H., B. Errea, J. Dukelsky, S. Pittel,and P. Van Isacker, Exactly solvable models of proton
and neutron interacting bosons, Phys. Rev. C 74 024314 1–7, (2011).
Chapter 10
Electromagnetic Transitions

Electric and magnetic multipoles. Transition probability and selection rules. Weisskopf units. Applica-
tions to nuclear spectra. The spin of the photon. 0+ → 0+ transitions. Internal and pair conversion.

In this Chapter we will study nuclear spectroscopic properties that can be extracted from experiments
using electromagnetic probes. We will assume that the nucleus is already excited in some initial state
and that it decays to a final state in the same nucleus through electromagnetic radiation by emitting a
photon. In nuclear electromagnetic decays the radiation, and the photon itself, is called ”gamma”. Our
aim is to analyze the characteristics of that photon. We will present the most important features of the
radiation field which are of interest for us without going into the details that can be found in standard
books on electromagnetism.
We thus starts by assuming that the electric field E and the magnetic field M in free space obey the
Maxwell equations, i. e.
1 ∂A
E=− (10.1)
c ∂t
M=∇×A (10.2)
where the vector field A is required to satisfy,

∂2 ∂2 ∂2 1 ∂2
( + + − )A = 0 (10.3)
∂x2 ∂y 2 ∂z 2 c2 ∂t2
with the gauge condition
∇·A=0 (10.4)
which shows that the electromagnetic fields are perpendicular to the direction of propagation. In quantum
mechanics this can be readily seen by noticing that Eq. (10.4) is the same as p · A = 0.
A solution of Eq. (10.3) is,
fk (r) = qk e−iωt Ak (r) (10.5)
where qk is a constant, k = ω/c and,

∇ × ∇ × Ak − k 2 Ak = 0 (10.6)

with the condition ∇ · Ak = 0.

We use j(r, t) and ρ(r, t) to denote the current and charge densities (operators) of the nucleus,

∂ i
∇ · j(r, t) = − ρ(r, t) = − [H, ρ(r, t)]. (10.7)
∂t ~
One may assume
A
X
ρ(r) = ei δ(r − ri ). (10.8)
i=1

10.1 The spin of the photon

We have seen in Chapter 1 that in a system with spherical symmetry the angular momentum is a
conserved quantity. We proved this by showing that the rotated wave function, which is a solution of the
Hamiltonian due to the spherical symmetry, is generated by applying the angular momentum operator
to the original wave function. From here we deduced that the angular momentum commutes with the
Hamiltonian. In the case of electromagnetism the field A which, as we showed above, determines the
10 Electromagnetic Transitions 108

y
y'

P
y'
x'
x'
y
θ
O x x

Figure 10.1: Relation between the systems of coordinates related by a rotation θ along the axis z = z 0 .

physical quantities, is a vector and, therefore, its rotation is a bit more complicated than in the scalar
case of the wave function.
We will rotate an angle θ along the z-axis. In the original system the coordinates of a point P are
P = (x, y, z) while in the rotated system we will use the notation P = (x0 , y 0 , z 0 ). As can be seen from
Fig. 10.1, the relation between the two coordinates is,
x0 = cos θx + sin θy
y0 = − sin θx + cos θy (10.9)
z0 = z
In the same fashion the components of the vector A transform as
A0x (x0 , y 0 , z 0 ) = cos θAx (cos θx + sin θy, − sin θx + cos θy, z)
+ sin θAy (cos θx + sin θy, − sin θx + cos θy, z)
A0y (x0 , y 0 , z 0 ) = − sin θAx (cos θx + sin θy, − sin θx + cos θy, z) (10.10)
+ cos θAy (cos θx + sin θy, − sin θx + cos θy, z)
A0z (x0 , y 0 , z 0 ) = Az (cos θx + sin θy, − sin θx + cos θy, z)
performing an infinitesimal rotation it is cos θ ≈ 1, sin θ ≈ θ and one gets,
Ax (cos θx + sin θy, − sin θx + cos θy, z) ≈ Ax (x, y, z) − iθLz Ax (10.11)
∂ ∂
where the definition of the angular momentum operator, i. e. Lz = − i(y ∂x x ∂y ) was used. Therefore
up to order θ it is
A0x (x0 , y 0 , z 0 ) = Ax (x, y, z) − iθ(Lz Ax + Ay (x, y, z)) (10.12)
repeating the same for the components A0y and A0z one gets, that
A0 = A − iθ(Lz + Sz )A (10.13)
where    
Ax −iAy
Sz  Ay  =  iAx  (10.14)
Az 0
It is left as an exercise to show that the eigenvalues of Sz are −1, 0 and +1. In the same fashion one can
define Sx and Sy by rotating along the axis x and y, respectively.
We thus see that the generator of rotations for a vector field is J = L + S, where S is the intrinsic
spin of the vector field with eigenvector S = 1. This implies that the photon carries a spin S = 1.
10 Electromagnetic Transitions 109

10.2 Electric and magnetic multipoles

The functions (10.5) form a complete set of vectors. Any electromagnetic field can be expressed as a
series within this representation. Each one of the components (10.5) corresponds to a quantized photon
with energy E = ~ω. They are determined by Eq. (10.6), which has two independent solutions. One of
them is given by
−i
AEλµ (kr) = ∇ × (r × ∇)(jλ (kr)Yλµ (θφ)) (10.15)
k
where jλ (kr) is a Bessel function and Yλµ (θφ) is the spherical harmonics eigenfunctions of the angular
momentum operator. This solution is called ”electric” component of the electromagnetic field. The other
solution is,
AM λµ (kr) = (r × ∇)(jλ (kr)Yλµ (θφ)) (10.16)
which is called the ”magnetic” component. In both cases the photon carries angular momentum λ with
z-projection µ. Therefore they are multipole components, with multipolarity λ,√of the electromagnetic
field. However, for the case λ = 0 there is no angular dependence (Y00 (θφ) = 1/ 4π) and since for any
angle independent function φ(r) it is ∇φ(r) ∝ r, the electromagnetic field vanishes (r × r=0). In other
words, there is no monopole photon in the electromagnetic field.
We have
∂
∇ × (r × ∇) = −∇ r + r∇2 , (10.17)
∂r
and
∇ · (r × ∇) = −∇ · (∇ × r) = −(∇ × ∇) · r = 0. (10.18)
For the parity we have πO = (−1)λ for Y λ , πO = −1 for the vectors r, ∇ and p, and πO = +1 for
pseudo vectors l = r × p and σ. We have

π̂Lπ̂ −1 = π̂r × (−i~∇)π̂ −1 = L, (10.19)

π̂∇π̂ −1 = −∇, (10.20)

and
1
π̂∇ × π̂ −1 = π̂ Ŝ · ∇π̂ −1 = −∇×, (10.21)
i~
where Ŝ is the spin operator mentioned above. The presence of the ∇ operator introduce an minus sign
under the pairty transformation. Thus the parity of the electromagnetic field is

−(−1)λ for electric multipole

λ
(10.22)
+(−1) for magnetic multipole

The two multipoles can be distinguished by the parity and are related to each other. We have
1
AE
λµ (kr) = ∇ × AM
λµ (kr). (10.23)
k
Integrating over the current density one obtains the electric component as,
−i(2λ + 1)!!
Z


O(Eλµ) = λ+1 jE (r) · ∇ × (r × ∇) jλ (kr)Yλµ (θφ) dr (10.24)
ck (λ + 1)
and the magnetic component becomes,
−(2λ + 1)!!
Z


O(M λµ) = jM (r) · (r × ∇) jλ (kr)Yλµ (θφ) dr, (10.25)
ck λ (λ + 1)
where the double factorial is defined as

(2λ + 1)!! = (2λ + 1)(2λ + 1 − 2)(2λ + 1 − 3)...3 · 1. (10.26)

The electric and magnetic current densities are defined as

1X pi pi
jE (r) = ei [δ(r − ri )] + δ(r − ri )], (10.27)
2 i M M
10 Electromagnetic Transitions 110

and
e~ X
jM (r) = µi ∇ × σi [δ(r − ri )]. (10.28)
2M i
In nuclear electromagnetic processes it is convenient to take into account the range of the quantities
involved. Thus nuclear photo transitions occur at large wavelengths of the photons as compared to the
nuclear radius R = 1.2A1/3 f m. Therefore kR is small (remember that the photon energy is E = hν =
hc/λ = k~c and in nuclear transitions it is E ≈ 1M eV ). One can expand the Bessel function as,

(kr)λ (kr)2
jλ (kr) = (1 − + ...). (10.29)
(2λ + 1)!! 2(2λ + 3)

We can take the first order of this expansion only in the long wave approximation. The electric multipole
can be rewritten as
 
i ∂
AEλµ (kr) = ∇ r(j λ (kr)Yλµ (θφ)) − r∇ 2
(j λ (kr)Yλµ (θφ))
k ∂r
  (10.30)
i ∂
= ∇ r(jλ (kr)Yλµ (θφ)) + k 2 r(jλ (kr)Yλµ (θφ)) .
k ∂r

The later term can be neglected in the long wave approximation. The multipole components of the field
acquire the simple forms,
−1
Z
O(Eλµ) = [H, ρ(r)]rλ Yλµ (θφ)dr
~kc
Z (10.31)
= ρ(r)rλ Yλµ (θφ)dr.

and
−1
Z
O(M λµ) = jM (r) · (r × ∇)rλ Yλµ (θφ)dr. (10.32)
c(λ + 1)
Selection rules are the results of conservation laws. They express certain symmetry conditions that
hold for the system under consideration. For instance, the invariance of a nuclear system as a whole
under spatial rotations leads to the conservation of the total angular momentum. When a nucleus
emits (absorbs) a photon, the initial (final) total nuclear angular momentum should be equal to the
sum of the final (initial) total nuclear angular momentum and the angular momentum carried by the
radiation. The selection rules are given by the triangle condition for the angular momenta ∆(Ji , Jf , λ).
The electromagnetic interaction conserves parity which depends on the spatial inversion properites of
the current densities. The operators for Eλ and M λ can be classified according to their transformation
under parity change:
P OP −1 = πO O (10.33)
For a given matrix element we have:

hΨf |O|Ψi i = hΨf |P −1 P OP −1 P |Ψi i = πi πf πO hΨf |O|Ψi i (10.34)

The matrix element will vanish unless πi πf πO = +1. Thus the transitions are divided into two classes,
the ones which do not change parity change πi πf = +1 which go by the operators with πO = +1:

πi πf = +1 for M1, E2, M3, E4... (10.35)

and the ones which do change parity change πi πf = −1 which go by the operators with πO = −1:

πi πf = −1 for E1, M2, E3, M4... (10.36)

The lowest allowed multipolarity in the decay rate dominates over the next higher one (when more
than one is allowed) by several orders of magnitude. The most common types of transitions are electric
dipole (E1), magnetic dipole (M1), and electric quadrupole (E2).
10 Electromagnetic Transitions 111

2 .2 6 1 c
E 2
1 .7 8 6 b
E 1
1 .5 7 9 a

E 2

+
0
1 5 0
E r 6 8

Figure 10.2: The first four levels in the nucleus 150 Er with the measured gamma ray transitions. The
task is to determine the spins and parities of the levels a, b and c.

10.3 Applications to nuclear spectra

The property that the electromagnetic transitions are strongly dependent upon the angular momentum
λ carried by the emitted photon is used to determine the angular momenta of the states involved in the
decaying process. From an experimental point of view the first thing that has to be done is to assess
the value of λ. This is done by measuring the angular distribution of the emitted radiation, which is
determined by the harmonic function Ylm (θφ). One measures also the polarization of the radiation and
thus whether the transition is of electric or magnetic character. The energy of the photon, that is of the
transition that emits the photon, is just hc/λ. The transition rate per second is obtained by measuring
the time that elapses between successive emitted photons, and from this one evaluates the corresponding
B-value and the half life, which is τ1/2 = ln 2 /T (σ, λ).
Once the multipolarities and energies of the possible transitions are known one assigns the angular
momenta of the states associated with those transitions following the probabilities given in Eq. (10.41).
This procedure can best be clarified by means of an example. We thus show in Fig. (10.2) the first four
states of the nucleus 150 Er (Z=68) with the corresponding gamma transitions. The ground state is 0+ ,
as it should be for an even-even nucleus. Since the state labeled a decays to the ground state emitting an
E2 photon, the spin Ja and parity πa of this state should be Jaπa = 2+ . The state a is connected to the
state b by an E1 transition, which implies that the parity of this state is minus. The angular momentum
should be 1 ≤ Jb ≤ 3. But it cannot be Jb = 1− because then it would decay to the ground state, and
not to the state a, by an E1 transition (since the large energy of Eb =1.786 MeV contributes with Eb3
to the transition rate. see Eq. (10.41)). Therefore the state b could be 2− or 3− . But in the B-factor,
Eq. (10.40) the matrix element is larger for states with aligned angular momenta, which implies that the
most likely assignment is Jbπb = 3− . For the same reason it is Jcπc = 5− .
In gamma transitions the emission of a λ=0 photon does not exist, as shown above by analyzing the
expression of the monopole component of the electromagnetic field. That was a mathematical proof. But
there is also a physical reason why such a photon does not exist, and that is that the intrinsic spin of the
photon is one.
10 Electromagnetic Transitions 112

10.4 Operators and transition rates

The interaction of the electromagnetic field with the nucleons can be expressed in terms of a sum of
electric and magnetic multipole operators with tensor rank λ
X
O= [O(Eλ)µ + O(M λ)µ ] (10.37)
λ,µ

The transition rate (transition probability) for a specific set of states and a given operator is given by:
  2λ+1
8π(λ + 1) k
Ti,f,λ = 2
B(i → f ) (10.38)
λ[(2λ + 1)!!] ~

where k is the wave-number for the electromagnetic transition of energy Eγ given by:

Eγ Eγ
k= = (10.39)
~c 197.33MeV fm
The last factor in Eq. (25.2) is referred to as a “reduced transition probability” B defined by:
2
X hJf ||O(λ)||Ji i
B(i → f ) = |hJf Mf |O(σλµ)|Ji Mi i|2 = (10.40)
2Ji + 1
µMf

This, which is also called ”B(σλ) value”, is a very important quantity because it contains all the nuclear
structure information. In particular, only are allowed transitions which satisfy the conservation laws
implicit in the matrix element hJf Mf |O(σλµ)|Ji Mi i. Thus, the B(σλ) value vanishes (and the corre-
sponding electromagnetic transition is forbidden) unless the parity condition πi πf = ±(−1)λ is fulfilled,
where πi (πf ) is the parity of the initial (final) state and the plus (minus) sign corresponds to electric
(magnetic) transitions. It should also fulfilled the angular momentum conservation law which, as usual,
is expressed by the triangular relation |Ji − Jf | ≤ λ ≤ Ji + Jf .
The B(σλ)-value also provides information on the structure of the initial and final states of the
decaying nucleus. Thus if, for instance, B(E1) is very large, as it is for transitions to the ground state
from what are called ”dipolar giant resonances”, then it indicates that the initial state of the nucleus is
built upon some uncommon mechanism induced by the nuclear dynamics.
It should be noticed that the unit of B(Eλ) is [B(Eλ)] = e2 f m2λ while it is [B(M λ)]= (µ/c)2 f m2λ−2 ,
where µ = eh/2mp is the nuclear magneton, and mp is the proton mass. Using these units the transition
rate per second becomes, for the most common values of the transfer angular momentum λ,

T (E1) = 1.59 × 1015 E 3 B(E1)

T (E2) = 1.22 × 109 E 5 B(E2) (10.41)
2 7
T (E3) = 5.67 × 10 E B(E3)
T (E4) = 1.69 × 10−4 E 9 B(E4)

T (M 1) = 1.76 × 1013 E 3 B(M 1)

T (M 2) = 1.35 × 107 E 5 B(M 2) (10.42)
0 7
T (M 3) = 6.28 × 10 E B(M 3)
T (M 4) = 1.87 × 10−6 E 9 B(M 4)

where the energies are in units of MeV.

The lifetime of the transition is given as 1/Ti,f,λ and the partial half-life is defined as

ln 2
tif
1/2 = . (10.43)
Ti,f,λ

The total half-life of a given (initial) state is given as

1 X 1
= . (10.44)
t1/2
f tif
1/2
10 Electromagnetic Transitions 113

With our definition of the reduced matrix element,

2 2
hJf ||O(λ)||Ji i = hJi ||O(λ)||Jf i (10.45)

B depends upon the direction of the transition by the factor of (2Ji + 1). For electromagnetic transitions
Ji is that for the higher-energy initial state. But in Coulomb excitation the initial usually taken as the
ground state, and one can use the notation B(↑) for this situation.

Electric transition operators

The electric transition operators are given by:

O(Eλ) = rλ Yµλ (r̂)etz e (10.46)

were Yµλ are the spherical harmonics. The etz are the electric charges for the proton and neutron in
units of e. For the free-nucleon charge we would take ep = 1 and en = 0, for the proton and neutron,
respectively. Although the bare operator acts upon the protons, we will keep the general expression in
terms of etz in order to incorporate the “effective charges” for the proton and neutron, which represent the
center-of-mass corrections and the average effects of the renormalization from wave function admixtures
outside the model space.
The most probable types of transitions are E1 and E2 (and M1 below). The E1 transition operator
is given with λ = 1 as r
(1) 3
O(E1) = rYµ (r̂)etz e = ret e (10.47)
4π z
The E2 transition operator is given with λ = 2:

O(E2) = r2 Yµ(2) (r̂)etz e (10.48)

Magnetic transition operators

The magnetic transition operators are given by:
" #
2gtlz
+ sgtz ∇ rλ Yµλ (r̂) µN
s
 
O(M λ) = l
λ+1
" # (10.49)
p  λ−1 λ 2gtlz  λ−1 λ s
= λ(2λ + 1) Y ⊗l µ + Y ⊗ s µ gtz rλ−1 µN .
λ+1

The M1 transition operator is given as

r
3  l
lgtz + sgtsz µN

O(M 1) = (10.50)
4π

One-body transition operator in the coupled form

As mentioned before, a one-body operator can be written as
X
Oµλ = hα|Oµλ |βia†α aβ , (10.51)
αβ

where the Greek letters denote the single-particle states |nljmi. We use j to denote the state |nlji. One
has,
X X
Oµλ = hα|Oµλ |βia†α aβ
jα jβ mα mβ
 
X X jα λ jβ
= hjα ||Oµλ ||jβ i (−1)jα −mα a†α aβ
−mα µ mβ (10.52)
jα jβ mα mβ

X [a†α ⊗ ãβ ]λµ

= hjα ||Oµλ ||jβ i √ ,
jα jβ
2λ + 1
10 Electromagnetic Transitions 114

where ãjm = (−1)j+m aj,−m . For the one-body transition density we have

hjf mf |[a†α ⊗ ãβ ]λµ |ji mi i = δαf δβi (−1)ji −mi < jf mf ji − mi |λµ > . (10.53)

The reduced single-particle matrix element for Eλ operator is given by:

r
la +λ+lb
ja +1/2 1 + (−1) (2ja + 1)(2λ + 1)(2jb + 1)
hja ||O(Eλ)||jb i =(−1)
 2  4π (10.54)
ja λ jb
× hja |rλ |jb ietz e
1/2 0 −1/2
The reduced single-particle matrix element for the spin part of the magnetic operator is:
p λ
hja ||O(M λ, s)||jb i = λ(2λ + 1)hja || Y λ−1 ⊗ s ||jb ihja |rλ1 |jb igtsz µN

 
p p  la 1/2 ja 
= λ(2λ + 1) (2ja + 1)(2ja + 1)(2λ + 1) lb 1/2 jb (10.55)
λ−1 1 λ
 

× hla ||Y λ−1 ||lb ihs||s||sihja |rλ−1 |jb igtsz µN

where p
hs||s||si = 3/2 (10.56)
The reduced single-particle matrix element for the orbital part of the magnetic operator is:
p
λ(2λ + 1) λ
hja || Y λ−1 ⊗ l ||jb ihja |rλ−1 |jb igtlz µN

hja ||O(M λ, l)||jb i =
λ+1
p
λ(2λ + 1) p
= (−1)la +1/2+jb +λ (2ja + 1)(2ja + 1) (10.57)
λ+1
 
la lb λ λ
hla || Y λ−1 ⊗ l ||lb ihja |rλ−1 |jb igtlz µN

×
jb ja 1/2
where
λ p
hla || Y λ−1 ⊗ l ||lb i =(−1)λ+la +lb (2λ + 1)lb (lb + 1)(2lb + 1)


λ−1 1 λ
 (10.58)
× hla ||Y λ−1 ||lb i
lb la lb
with r  
λ−1 la (2la + 1)(2lb + 1)(2λ − 1) la λ−1 lb
hla ||Y ||lb i = (−1) (10.59)
4π 0 0 0
For the M1 operator the radial matrix element is:
hja |r0 |jb i = δna ,nb (10.60)
and the reduced single-particle matrix element simplify to:
r
3
hja ||O(M 1, s)||jb i = hja ||s||jb iδna ,nb gtsz µN
4π
r
3 p
= (−1)la +ja +3/2 (2ja + 1)(2jb + 1) (10.61)
4π
 
1/2 1/2 1
× hs||s||siδla ,lb δna ,nb gtsz µN
jb j a la
and
r
3
hja ||O(M 1, l)||jb i = hja ||l||jb iδna ,nb gtlz µN
4π
r
3 p
= (−1)la +jb +3/2 (2ja + 1)(2jb + 1) (10.62)
4π
 
la lb 1
× hla ||l||lb iδna ,nb gtlz µN
jb ja 1/2
10 Electromagnetic Transitions 115

where p
hla ||l||lb i = δla ,lb la (la + 1)(2la + 1) (10.63)
Thus the M1 operator can connect only a very limited set of orbits, namely those which have the same
n and l values.
We also have p
hja ||j||jb i = δa,b ja (ja + 1)(2ja + 1), (10.64)
and p
hs||s||si = s(s + 1)(2s + 1). (10.65)

Moments in terms of electromagnetic operators

There are two type of electromagnetic moments: the electric moments associated with the static distribu-
tion of charge and the magnetic moments associated with the magnetic currents. The parity conservation
requires λ=even for the static moments and λ=odd for the magnetic moments. The electromagnetic
moment for λ=0 simply counts the number of particles with charge ek .
A compilation of experimental magnetic and quadrupole moments
http://www.nndc.bnl.gov/publications/preprints/nuclear-moments.pdf
The operator for electromagnetic moment can be expressed in terms of the electromagnetic transition
operators. By the parity selection rule, the moments are nonzero only for M1, E2, M3., E4,.... The most
common are:
r
4π
µ= hJ, M = J|O(M 1)|J, M = Ji
3
r   (10.66)
4π J 1 J
= hJ||O(M 1)||Ji
3 −J 0 J

and
r
16π
Q= hJ, M = J|O(E2)|J, M = Ji
5
r   (10.67)
16π J 2 J
= hJ||O(E2)||Ji
5 −J 0 J

10.5 Nuclear matrix elements

The matrix element of a one-body operator can be written as
X
hJf Mf |O(λµ)|Ji Mi i = hα|O(λµ)|βihJf Mf |c†α cβ |Ji Mi i (10.68)
αβ

Electromagnetic transitions and moments (and beta decays) depend upon the reduced nuclear matrix
elements hf ||O(λ)||ii. These can be expressed as a sum over reduced one-body transition densities times
single-particle matrix elements:
X
hf ||O(λ)||ii = T (jα jβ λ)hjα ||O(λ)||jβ i (10.69)
jα jβ

where the reduced one-body transition density is given by

λ
hf || αj†α ⊗ α̃jβ ||ii

T (jα jβ λ) = √ (10.70)
2λ + 1
The labels i and f are a short-hand notation for the initial and final state quantum numbers. Thus
the problem is divided into two parts, one involving the nuclear structure dependent reduced one-body
transition density, and the other involving the reduced single-particle matrix elements.
10 Electromagnetic Transitions 116

For a closed shell plus one particle the only term contributing to the sum comes from the transition
between two specific particle states as
λ √
hjf || αj†α ⊗ α̃jβ ||ji i = δαf δβi 2λ + 1,

(10.71)

and
hJf = jf ||O(λ)||Ji = ji i = hjf ||O(λ)||ji i. (10.72)
The reduced transition probability for this cases is:
2
hjf ||O(λ)||ji i
B(λ) = (10.73)
2ji + 1

Single-j configurations
For a closed shell plus n particles in an orbital j the matrix elements reduce to:

hj n , αf , Jf ||O(λ)||j n , αi , Ji i = T (jλ)hj||O(λ)||ji (10.74)

q X
T (jλ) =n (2Jf + 1)(2Ji + 1) (−1)Jf +λ+J+j
αJ
  (10.75)
Ji Jf λ n−1 n n−1 n
× [j (αJ)jJf |}j αf Jf ][j (αJ)jJi |}j αi Ji ],
j j J
where the square brackets denote the one-particle coecients of fractional parentage (cfp’s).
For n = 2, the cfp’s are unity and the sum only contains J = j in which case we have
 
Ji Jf λ
q
T (jλ) = (−1)Jf +λ+1 n (2Jf + 1)(2Ji + 1) (10.76)
j j j

The reduced transition rate becomes:

 2
Ji
jf λ 2
B(λ) =n2 (2Jf + 1) hj||O(λ)||ji
j
j j
2 2 (10.77)
hj||O(λ)||ji

2 J i jf λ
=n (2Jf + 1)(2j + 1) .
j j j 2j + 1

0+ → 0+ transitions. Internal and pair conversion

The physical reason why there is no electromagnetic decay with the photon carrying angular momentum
λ = 0 is that the spin of the photon is unity. In general this implies that the decay proceeds through
other angular momenta. For instance if both the initial and final states have angular momenta J, then
it should be, since λ = 0 is not allowed, 1 ≤ λ ≤ 2J. However, if J = 0 then the electromagnetic decay
is not possible. In this case there are two different mechanism which can induce the decay.
Assuming that the decay proceeds from an excited state 0+ +
2 to a lower state 01 , one of the mechanisms
of decay is called ”internal conversion”. In this the nucleus transfers energy from the state 0+ 2 and decays
to the state 0+1 . This energy is absorbed by the electrons that surrounds the nucleus. The most affected
of those electrons are the ones which are closer to the nucleus, that is the ones that move in l = 0 orbits,
and of these the closest to the nucleus are the ones with lowest energy, i. e. the ones called ”K electrons”.
The internal electron thus affected converts the energy transferred from the nucleus (therefore the name
”internal conversion”) into kinetic energy and escapes the atom as a free electron. The nucleus performs
the transition 0+ +
2 → 01 .
The second mechanism by which the nucleus performs that transition requires that the energy between
the states 0+ +
2 and 01 is larger than twice the electron mass me = 0.511 MeV, i. e. larger than 1.022 MeV.
If this happens then there is enough energy to create an electron-positron pair. Any energy in excess of
2me is transformed in kinetic energy which is shared by the electron-positron pair. Therefore this form
of decay is called ”pair conversion”.
10 Electromagnetic Transitions 117

10.6 Weisskopf units

One sees that the decay rates (10.41) are very much dependent upon the multipolarity of the emitted
photon, decreasing as λ increases. This property is only determined by the electromagnetic field since
the nuclear dynamics is contained in the B-values. As mentioned above, this quantity depends upon the
structure of the nucleus. A rough estimation for the most simple case, that is the one corresponding to
the decay from a single-particle state ji to a final one jf , is obtained by using Eqs. (10.31) and (10.32).
For this one assumes first that the radial wave functions are constant inside the nuclear radius and vanish
outside it. Second the angular part of the matrix element is evaluated assuming that the transition occurs
from the state Ji = λ + 1/2 to the state jf = 1/2. And third in the magnetic case (10.32) appears the
quantity F = λ2 (gs − 2(λ + 1)−1 gl )2 , where gl and gs are the g-factors of Chapter 2. This quantity is
assumed to have the value F =10. With these approximations the B-values become,

1.22λ 3 2 2λ/3 2
BW (Eλ) = ( ) A e f m2λ , (10.78)
4π λ + 3
and
10 2λ−2 3 2 (2λ−2)/3 e~
BW (M λ) = 1.2 ( ) A f m2λ−2 . (10.79)
π λ+3 2M c
For instance we have
e~
BW (M 1) = 1.790 . (10.80)
2M c
These values allow one to estimate whether the transition probabilities are strongly dependent upon λ
even when the influence of the nuclear structure contained in the B-values are taken into account. For
this we evaluate the ratio R = BW (σ, λ)/BW (σ, λ + 1), which turns out to be independent of σ, i. e. it
is the same for electric or magnetic components.
BW (σ, λ) 1 λ + 4
2
R= = (10.81)
BW (σ, λ + 1) 1.44A2/3 λ + 3

which for a light nucleus with A = 20 is R ≈ 0.1((λ + 3)/(λ + 4))2 while for A = 200 R ≈ 0.02((λ +
3)/(λ + 4))2 . These numbers do not affect the conclusion that the transition probabilities (10.41) are
strongly dependent upon λ. Therefore a nucleus in an initial state with a given angular momentum Ji
will decay most probably by emitting a photon carrying the minimum angular momentum λ allowed by
the selection rules.
The importance of the Weisskopf units, Eqs. (10.78) and (10.79), is that they allow one to estimate
the strength of the B-values. In a transition where only one state takes part, then the B-value should be
≈ 1 Weisskopf unit. This will be the case of a particle moving outside a double magic core. But, as we
saw in the last Chapter, a nuclear excited state can consist of many configurations, and therefore many
single-particle states can contribute to the B-value. If these contributions occur all with the same phase
then the B-value will become very large in Weisskopf units. Since in this case the single-particle states
add collectively to the B-value and, therefore, to the transition probability, the decaying state is called
”collective state”. A typical collective state is the giant dipole resonance mentioned above, for which the
B(E1)-value for the decay to the ground state can be greater than 100 Weisskopf units.
In some cases the reduced transition probability B(Eλ) is expressed in the unit of eλ barnλ . We have
1barn=100fm2 and eλ barnλ =102λ eλ fm2λ . For instance we have

BW (E1) = 6.446 ∗ 10−4 A2/3 e ∗ barn = 6.446 ∗ 10−2 A2/3 e ∗ f m, (10.82)

and
BW (E2) = 5.940 ∗ 10−6 A4/3 e2 barn2 = 5.940 ∗ 10−2 A4/3 e2 f m2 . (10.83)

10.7 Deuteron Structure

The deuteron is the only bound state of 2 nucleons, with isospin T = 0, spin-parity J π = 1+ , and binding
energy EB = 2.225MeV. For two spin 1/2 nucleons, only total spins S = 0, 1 are allowed. Then the
orbital angular momentum is restricted to J − 1 < l < J + 1, i.e., l = 0, 1 or 2. Since the parity is
π = (−)l = +, only l = 0 and l = 2 are allowed; this also implies that we have S = 1.
10 Electromagnetic Transitions 118

Figure 10.3: Radial wave functions R0 (r) and R2 (r) calculated from realistic N N interactions.

If the Hamiltonian is
~2 1 d2 ~2 L2
H=− r + + VC (r) + VT (r)S12 (10.84)
M r dr2 M r2
using the following relation,
M
√ M M
√ M M
S12 Y001 = 8Y211 ; S12 Y211 = 8Y011 − 2Y211 ; L2 Y011
M
= 0; L2 Y211
M M
= 6Y211 (10.85)
we find the radial equations
~2 d2 √
 
+ E − V c (r) uS = 8VT (r)uD
M dr2
√
 2 2  
~ d 6
− 2 + E + 2VT (r) − Vc (r) uD = 8VT (r)uS (10.86)
M dr2 r
These equations can be solved numerically.
Other important information on the structure of the deuteron comes from the values of the magnetic
moment µ and quadrupole moment Q:
µ = 0.8574µN ; Q = 0.2857e fm2 (10.87)
Since Q 6= 0, the deuteron cannot be pure l = 0. But generally l = 0 is energetically favored for a central
potential. Therefore, we write the deuteron wave function as a linear combination of S- and D- waves
ψ(r) = aψ3 S1 (r) + bψ3 D1 (r)
1 1
= aR0 (r)Y011 + bR2 (r)Y211 (10.88)
√
where a and b are constants with a2 + b2 = 1. R0 and R2 are the radial wave functions.
1
Y011 = Y00 χ00 (10.89)
X
1
Y211 = h1(1 − M )2M |11iY2M χ1,1−M
M
r r r
1 3 3
= Y20 χ1,1 − Y21 χ1,0 + Y22 χ1,−1 (10.90)
10 10 5

Magnetic Moment
As mentioned before, the free-nucleon values for the g-factors are gpl = 1, gnl = 0, gps = 5.586 and
gns = −3.826. The magnetic moment operator can be rewritten as
X
µ = µN (gs szi + gl lzi ) (10.91)
i
10 Electromagnetic Transitions 119

where gs = 4.706τi +0.88, where the first term is isovector, and the second term is isoscalar. gl = (τi +1)/2.
Since the deuteron is an isoscalar particle with T = 0, only the isoscalar magnetic moment operator
contributes to µ. Then, the above equation becomes,
X 1
µ = µN (0.88szi + lzi )
i
2
2
X 
z 1 z
= µN 0.88hsi iM =1 + hli iM =1
i=1
2
 
1
= µN 0.88hS z i + hLz i
2
The matrix element of Sz is calculated to be
1 1
hY011 |Sz |Y011 i = 1
1 1
hY211 |Sz |Y011 i = 0
1 1
X 2
hY211 |Sz |Y211 i = h2(1 − MS )1MS |11i = −1/2 (10.92)
MS

Thus, for pure l = 0 or l = 2 states we would have the values µ = 0.88µN , 0.31µN . More generally we
obtain the relation
µ = a2 (0.88) + b2 (0.31) µ0 = (0.88 − 0.57b2 )µ0
 
(10.93)
Therefore, the experimental value µD = 0.857µN implies that b2 = 0.04. However, in more sophisticated
treatments one finds that it is quantitatively important to explicitly include the effects of meson exchanges
on the magnetic moment. For example, the virtual photon can couple to the pion in flight between the
nucleons and convert it to a ρ meson. (These effects are clearly observed in elastic electron deuteron
scattering at higher momentum transfers, as discussed below.) Including the effects of these ”meson-
exchange” currents on the deuteron’s magnetic moment yields values of b2 = 0.05 ∼ 0.07.

Quadrupole Moment
The quadrupole moment of the deuteron is calculated to be
r Z " 2 #
16π ∗
X
Q= ψJ=M =1 (r) etz ri Y20 (r̂i ) ψJ=M =1 (r)d3 r
2
5 i=1
r (10.94)
r2
Z
16π ∗
=e ψJ=M =1 (r) Y20 (r̂)ψJ=M =1 (r)d3 r,
5 4
where we have used the fact that for each nucleon the distance from the center of mass is only half the
distance between them, ri = r/2. Inserting the wave function introduced above, we get
r ( Z Z Z
π ∗
Q =e |a2 | r2 R0 (r)2 dr Y00 Y20 Y00 dΩ + 2Re(ab∗ ) r2 R0 (r)R2 (r)dr
5
X Z
∗
× h1(1 − M )2M |11i Y00 Y20 Y2M dΩ (10.95)
M
Z Z )
2 2 2
X 2 ∗
+ |b| r R2 (r)dr × h1(1 − M )2M |11i Y2M Y20 Y2M dΩ .
M

After evaluating the angular integrals and putting in the CG coefficients, one finds
(√ )
|b|2
Z Z
2 ∗ 2 2 2
Q=e Re(ab ) drr R0 (r)R2 (r) − drr R2 (r) (10.96)
10 20

For our purposes, we will use our knowledge that b = 0.2  1 from the magnetic moment analysis
and keep only the first term. This will give us an approximate expression that we can set equal to the
experimental value Qexp = 0.286e fm2 to obtain the result
√ Z
0.2 2
Q'e r4 R0 R2 dr = 0.286e fm2 (10.97)
10
10 Electromagnetic Transitions 120

Solving for the unknown radial integral yields

Z
r4 R0 R2 dr ' 10.1fm2 (10.98)

for the radial integral. This value seems Rquite reasonable given that the mean squared charge radius of
the deuteron is 4.0 fm2 . [Note: hrch
2
= 14 R02 r4 dri]
The dominant S-D interference term in the quadrupole moment has MS = 1. So the spins of both
the two nucleons are predominantly aligned parallel to ẑ. Let’s simply take σ1 = σ2 = +ẑ, and then
σ1 · σ2 = +1. Then we need to consider the relative orientation of r̂, and we will focus (see Fig. 10.7.2
on two extreme cases: (a)r̂||ẑ and (b)r̂⊥ẑ.

Figure 10.4: Schematic picture of the relative orientation of ẑ and r̂.

In case (a)σ1 · r̂ = σ2 · r̂ = 1, so we have S12 = +2 for this geometrical arrangement. This is a prolate
configuration so we expect Q > 0 for case (a). In case (b) we have σ1 · r̂ = σ2 · r̂ = 0 so S12 = −1 and
the oblate shape relative to the ẑ axis would imply Q < 0.
Since experimentally Q > 0, case (a) must be energetically favored which corresponds to VT (r) < 0.
This then gives an attractive force when the configuration is such that S12 > 0 (case (a)) and a repulsive
force when S12 < 0(case (b)).
One may also read
http://www.nature.com/nature/journal/v144/n3645/abs/144476a0.html
10 Electromagnetic Transitions 121

10.8 Homework problems

Exercise 1:
a) Show that the electric component AE λµ (r) of the electromagnetic field carries parity (−1)
λ+1
and
M λ
that for the magnetic component Aλµ (r) the parity is (−1) .
b) Which are the parity and triangular conditions that the B(σλ)-value should fulfilled?

Exercise 2:
The first excited state of the nucleus 116 Sn50 lies at 1.294 MeV and decays to the ground state by
emitting an E2 photon with a half life τ1/2 = 0.374 ps (1 ps =10−12 sec).
a) Determines the corresponding B(E2)-value in units of e2 f m4 .
b) As a) but in Weisskopf units.

Exercise 3:
a) As Exercise 2 but for the level 210 Pb82 (2+
1 ) lying at 800 keV. It decays by an E2 transition to the
ground state with a half live of 17 ps.
b) What conclusion you draw from the differences of the B(E2) values in Weisskopf units for the states
116
Sn(2+
1 ) and
210
Pb(2+
1) ?

Exercise 4:
The spectrum of 90 Zr40 consists of the 0+ ground state and the excited states 0+ +
2 (at 1.761 MeV), 21
− − + + +
(2.186 MeV), 51 (2.319 MeV), 41 (2.739 MeV), 41 (3.077 MeV), 61 (3.448 MeV) and 81 (3.589 MeV).
Determine the most likely decay pattern of the excited states.

Exercise 5:
Show that the photon has intrinsic spin S=1?
 Bibliography

[1] P. J. Brussaard and P. W. M. Glaudemans, Shell-model applications in nuclear spectroscopy, North-

Holland Pub. Co. (Amsterdam ; New York; Oxford), 1977
[2] B. Alex Brown, Lecture Notes in Nuclear Structure Physics, 2005
Chapter 11
Beta and alpha readioactive decays

Beta decay. Q-values in β − , β + and electron capture decays. Theory of beta decay. Fermi and Gamow-
Teller operators. The logaritm of f t. Alpha decay. Thomas expression for the width.

We have seen in the previous Chapter that a nucleus in an excited state can decay to a lower lying state
in the same nucleus by emitting a gamma ray. There are other paths which the nucleons in an unstable
nucleus can take to eventually reach a more stable configuration. These paths are called ”channels”.
Some of these channels may require that particles other than photons are emitted. In this Chapter we
will first study channels in which electrons are emitted. These are called beta decay channels. In the
second part we will study channels where alpha particles, i. e. 4 He2 nuclei, are emitted. These are called
alpha decay channels.
The names of these three forms of radiation, i. e. ”alpha”, ”beta” and ”gamma”, have an historical
origin. The alpha and beta modes of radioactive decay were the first to be discovered. Already in 1896
Becquerel found that charged particles were emitted by uranium, and soon afterwards Marie and Pierre
Curie confirmed these observations. In 1899 Rutherford discovered that the radiation could be divided in
two separated strings according to the penetration power they had. He called this two forms of radiation
”alpha” and ”beta”. Alpha radiation could penetrate only thin foils of aluminium while beta rays could
go through many millimeters of the same material. Only in 1900 the third form of radiation was found,
which had even more penetration power than beta rays, and these were called ”gamma” rays. It was
soon found that the alpha ray was actually an alpha particle, that the beta ray was an e− electron and
that the gamma ray was a photon.
In all these decays the outgoing particle should travel with a kinetic energy which is just the diference
between the energies of the initial and final states. However, in 1911 Lise Meitner and Otto Hahn found
that the electron emitted in beta decay had a continuum spectrum, thus seemingly violating the energy
conservation law. It took a long time before the explanation of this phenomenom was found. It was Pauli
who in 1930 proposed that a particle that interacts very weakly with matter (and therefore undetected up
to that time) was also emitted together with the electron. This hypothetical particle was called neutrino.
The total energy in the beta decay process was thus shared between the electron and the neutrino. Since
then the neutrino was measured experimentally and was found to play a crucial role in many branches
of physics, particularly in beta decay processes. These are the processes which we will now analyze.

11.1 Introduction to beta decay

The lighest nucleus that decays by beta emission is the free neutron, which undergoes the spontaneous
process
n → p + e− + ν̄ (11.1)
with a half life of about 10.3 minutes. In this equation n (p) is a neutron (proton), e− is an electron
and ν̄ is an antineutrino. Antiparticles have the same mass and spin as the corresponding particles, but
opposite charge and spin projection. Thus the antiparticle of the electron (e− ) is the positron (e+ ). The
neutron is heavier than the proton, which explains why it is unstable.
The neutrino ν is a chargeless particle which interacts very weakly with other particles. Therefore
the interaction responsible for beta decay is called ”weak interaction”. One can assess its intensity by
comparing the dimensionless strengths of the forces known in Nature. Thus, for the nuclear interaction
this strength is approximately 10, for the electromagnetic field it is 1/137, for the weak interaction 10−23
and for the gravitational field 10−45 .
There are experimental evidences showing that the neutrino has a very small mass, probably less than
1 eV. However, for the purpose of this Course we will assume that the neutrino is massles and, therefore,
that it moves at the speed of light at a certain frequency ν and energy hν.
11 Beta and alpha readioactive decays 124

In all decay processes the total charge is conserved. Therefore the decay of the neutron to a proton
requires that an electron e− is also emitted. This is why the name β − is used for this decay.
The electron e− and the antineutrino ν̄ which are emitted are called ”leptons”. Even the number of
leptons is conserved in all decay processes. The application of this law requires that one has to give a
value to the lepton number. For convention the electron and the neutrino have lepton number 1, while
the positron e+ and the antineutrino ν̄ have lepton number -1. With this convention in the decay (11.1)
the lepton number is zero before as well as after the decay.
The electron and the antineutrino appear always together in beta decay. This pair of leptons form
what is called the electronic leptons or the first ”lepton generation”. There are three lepton generations.
Besides the electron and its neutrino, also called electron neutrino and denoted by νe , there is the muonic
leptons, comprising a particle called muon (µ) and its muon neutrino (νµ ) and the tauonic leptons,
comprising a particle called tau (τ ) and the corresponding tau neutrino (ντ ). However, we will only deal
with the first lepton generation. That is our leptons will only be the electron, the electron neutrino and
their antiparticles.

Q-values in β − , β + and electron capture decays

The decay energy associated to the process (11.1), which is called ”the Q-value”, is given by
Qβ − = mn c2 − mp c2 − me c2 (11.2)
Since mn c2 =939.566 MeV, mp c2 =938.272 MeV and me c2 =0.511 MeV, the Q-value is Qβ − =0.783 MeV.
This is the kinetic energy shared by the three particles in the exit channel of Eq. (11.1).
For the β + decay, i. e.
p → n + e+ + ν (11.3)
it is a positron and a neutrino that are emitted. Since the masses of the antiparticles are the same as the
ones of the corresponding particles one has, Qβ + = mp c2 − mn c2 − me c2 =-1.805 MeV, which implies that
this decay, with negative kinetic energies for the outgoing particles, is forbidden. Or, in other words, for
the process (11.3) to take place it is necessary that energy is provided from the outside.
Another form of decay is electron capture (EC) by a proton in the process,
p + e− → n + ν (11.4)
2 2 2
for which the Q-value is QEC = mp c + me c − mn c −=-0.783 MeV, which is also a forbidden transition.
All these transitions can also occur within the nucleus, where the Q-value of the decay from a particular
nucleon is modified by the energies provided by the other nucleons. Thus, if the decay occurs from a
nucleus (Z,N), where the nucleon number is A=N+Z, the nuclear beta minus decay becomes
(Z, N ) → (Z + 1, N − 1) + e− + ν̄ (11.5)
with the Q-value given by Qβ − = M (Z, N ) − M (Z + 1, N − 1) − me c2 , where M is the atomic mass
(in units of energy) given by the sum of the masses of the neutrons, protons and electrons, minus the
corresponding binding energies, i. e.
M (Z, N ) = N mn c2 + Zmp c2 + Zme c2 − Bel (Z, N ) − B(Z, N ) (11.6)
where Bel (Z, N ) is the binding enery of the electrons in the atom and B(Z, N ) is the nuclear binding
energy. We have assumed that the atom is not ionized and therefore the number of electrons is the same
as the number Z of protons.
The value of Qβ − can now be positive or negative, depending upon the structure of the nuclei involved
in the decay.
For the nuclear β + decay it is,
(Z, N ) → (Z − 1, N + 1) + e+ + ν (11.7)
2
and Qβ + = M (Z, N ) − M (Z − 1, N + 1) − me c , which can also be positive or negative according to the
structure of the nuclei.
Finally, the nuclear electron capture occurs when one of the electrons orbiting around the nucleus
is absorved by a proton in the nucleus. The proton is transformed into a neutron and an electrino is
emitted following the process (11.4), i. e.
(Z, N ) + e− → (Z − 1, N + 1) + ν (11.8)
2
with QEC = M (Z, N ) + me c − M (Z − 1, N + 1)
11 Beta and alpha readioactive decays 125

Beta decay rates and the ft value

A formalism was developed during the 1960’s and early 1970’s, called electroweak theory, which describes
in an unified fashion the electromagnetic and the weak interactions. The description of decay processes
within this formalism requires a knowledge of field theory which is outside the scope of this Course.
However, at the energies that we will be concerned a simpler form of a field theory can explain well the
properties which are relevant for our purpose. This is actually the first field theory ever applied outside
electromagnetism. It was performed by Fermi, as we show below.
The creation of the neutrino in beta decay implies that the force that induces the transition from
the initial to the final state is weak since, as mentioned above, the neutrino interacts very weakly with
matter.
Neither the electron nor the neutrino exist in the nucleus, therefore they must be formed at the
moment of their emission, just as a photon is created at the moment of its emission in electromagnetic
decay. From this simple argument Fermi developed a theory of beta decay by following the same steps as
in electromagnetic decay. We will follow Fermi’s argument, which is very simple and forms the ground
on which all field theories are based.
We will assume, for simplicity, that the Coulomb force does not affect the escaping electron, which
would be true only is the electron energy is large. This approximation will be lifted later. With this
assumption the lepton-lepton and lepton-nucleon interaction is weak and therefore one can consider
that the leptons move freely inside the nucleus. One can thus consider that the lepton wave functions
inside the nucleus are plane waves of the form Ψe (r) = Ne eije ·r for the electron and for the neutrino
Ψν (r) = Nν eijν ·r , where k is the wave number, k = 2π/λ, and N is a normalization constant.
The emission probability will depend upon the expectation value for the leptons to be inside the
nucleus. With r within the nucleus this expectation value is |Ψe (r)|2 |Ψν (r)|2 = |Ne |2 |Nν |2 . This describes
the lepton sector of the decay. For the nucleon sector, Fermi compared with electromagnetic transitions
and concluded that another factor that has to appear in the beta emission probability is the matrix
element of some operator between the initial and final nuclear states, in the same fashion as the matrix
element of the multipole operator M determines the transition probability per second, called T in Eq.
(10) of the previous Chapter. We will assume that it is a neutron that decays and that this neutron
belongs to a nucleus in a state |α(N )i, as the nucleus called (Z, N ) in Eq. (11.5). The nucleus in the
final state, that is (Z + 1, N − 1), will be assumed to be in the state |β(P )i. The nuclear matrix element
should thus have the form Z
M = Ψ∗β(P ) (r)OΨα(N ) (r)dr (11.9)

where O is some operator that one has to introduce to explain the experimental data.
The transition probability per unit of time is given by the ”golden rule” of quantum mechnics, i. e.
it is the square of the transition matrix element times the density of states, which in our case is,
2π dN
T = |Ne Nν M|2 (11.10)
~ dE
√
We will assume that the system is contained in a box of volume V and, therefore, Ne = Nν = 1/ V .
dN /dE is the density of final states. A change in E implies a change in the energy of the electron, which
determines any change in the energy of the neutrino. This is because the transition energy Et = Ee + Eν
is a constant, i. e. dEt = 0, and one has dE = dEe = −dEν . One can then write the density of states as
dN /dE = dN /dEe .
To calculate dN /dEe we start by evaluating the number of states with momenta between p and p+dp.
One of the Exercises corresponding to this Chapter is to evaluate that number for the plane wave confined
within the box and for the particle i (i is electron or neutrino). It is

p2i dpi V
dNi = (11.11)
2π 2 ~3
and the total number of states is
p2e dpe V p2ν dpν V
dN = dNe dNν = (11.12)
2π 2 ~3 2π 2 ~3
For the neutrino (massless) one has dEν = cdpν . The transition probability has to be a positive number,
therefore we replace dEν = |dEe | = cdpν . The maximum value of the electron energy Ee corresponds
11 Beta and alpha readioactive decays 126

to Eν = 0. Therefore the transition energy is Et = Eemax and Eν = cpν = Eemax − Ee . The neutrino
is a relativistic particle and even the electron is emitted at high speed. Therefore one has to use the
relativistic expression for the energy E, which for a particle with mass m carrying a linear momentum p
is E 2 = m2 c4 + c2 p2 . Therefore it is p2e p2ν = (Ee2 /c2 − me c2 )(Eemax − Ee )2 /c2 and
dN V2
= (E 2 /c2 − me c2 )(Eemax − Ee )2 (11.13)
dEe dpe 4π 4 ~6 c3 e
which gives an excellent representation of the shape of the beta decay spectrum that Fermi wanted to
explain, supporting his theory as well as the neutrino assumption.
The transition probability (11.10) for the electron in a range of momentum dpe becomes,
|M|2
T dpe = (E max − Ee )2 p2e dpe (11.14)
2π 3 ~7 c3 e
notice that the volumen V has been cancelled out in this expression, p which implies that its exact value
is irrelevant, as it should be. Calling Pe = pmax
e one has Eemax = m2e c4 + c2 Pe2 . Replacing this in Eq.
(11.14) one obtains,
|M|2 p 2 4 p
2
T dpe = m e c + c 2P 2 −
e m2e c4 + c2 p2e p2e dpe (11.15)
2π 3 ~7 c3
The total transition probability Ttotal per unit time is obtained by integrating this equation over the
linear momentum pe . The mean life τ = 1/Ttotal is obtained from,
Z Pe p
1 |M|2 p
2
= 3 7 3
m2e c4 + c2 Pe2 − m2e c4 + c2 p2e p2e dpe (11.16)
τ 2π ~ c 0
To obtain this equation we assumed that the electron was not affected by the Coulomb interaction and
also that the decaying neutron existed in the nucleus as a free particle. None of these two assumptions
is valid in general. The inclusion of the Coulomb interaction is a relatively easy task, but to assess the
extend to which the nucleon is free can be a rather challenging task. The outcome of this calculation is
that in Eq. (11.16) a Fermi factor F (Zf , pe ) has to be added. The half life τ1/2 = ln 2 τ thus becomes,
Z Pe
|M|2 p p
2 −1
m2e c4 + c2 p2e p2e dpe

τ1/2 = F (Z , p ) m2 c4 + c2 P 2 − (11.17)
f e e e
2 ln 2π 3 ~7 c3 0
One sees that the transition energy Et = Eemax , or its corresponding momentum Pe . appears only in the
factor Z Pe
5 7
p p
2
f (Pe )me c = F (Zf , pe ) m2e c4 + c2 Pe2 − m2e c4 + c2 p2e p2e dpe (11.18)
0
and the quantity
2 ln 2π 3 ~7
f τ1/2 = (11.19)
|M|2 m5e c4
is a constant independent of the energy. This is called f t factor. One determines experimentally this
quantity by evaluating the function f (Pe ) numerically and using the experimental half life. We will see
below that the f t factor provides a fair measure of the beta decay transition probability.
We have
2π 3 ~7
= 1.8844 × 10−94 erg2 cm6 s. (11.20)
m5e c4
We have analyzed β − decay only, but the formalism is the same for β + decay. In the case of electron
capture the nucleus traps an electron of the atomic electronic cloud. For this to happen there should be
a large overlap between the electron and the nuclear wave functions. The most internal electronic orbit
corresponds to K-electrons, i. e. electrons moving in the lowest atomic orbit. The electron thus absorved
undergoes the beta-like decay process shown in Eq. (11.4). However there is an important difference since
the K-electron is absorved at the energy of the K-shell. Instead, as we have seen above, in beta decay
the electron is emitted into the continuum with any energy, and the corresponding transition probability
has to take into account the density of continuum states. Therefore, in electron capture the statistical
factor (11.18), which is an outcome of the density of states, does not appear.
Another difference between electron capture and beta decay is that in electron capture the K-electron
moves in a l = 0 state, and cannot be considered a plane wave. It is not difficult to include the proper
electron wave function in the transition matrix element, but we will not go into more details of this
process since conceptually it is similar to beta decay.
11 Beta and alpha readioactive decays 127

11.2 Fermi Theory of allowed beta decays

One has still to evaluate the matrix element M, and to do that one has first to choose the operator
O in Eq. (11.9). This is a crucial point for the Fermi theory as well as for all field theories developed
afterwards. There are some restrictions which confine the choice one can make of O. The most important
of these restrictions is that the system is relativistic and, therefore, the physics should be the same for
observes moving at different (but constant) relative speeds while keeping the light velocity c constant.
Or, more precisely, the system should be invariant under Lorentz transformations. The types of beta
decay can be classified by the angular momenta carried away by the electron and neutrino. The most
important are those for ∆l = 0 which are referred to as “allowed” beta decay. There are two type of
allowed beta decay – Fermi (F) and Gamow- Teller (GT).
Fermi took first the simplest form, which is just O = 1, resulting in what is known as the Fermi
matrix element, i. e. Z
MF = h|O|i = gV Ψ∗β(P ) (r)Ψα(N ) (r)dr (11.21)

where gV is the strength of the coupling that gives rise to the emission, as the fine structure constant
α = e2 /(~c) ≈ 1/137 is the strength of the electromagnetic interaction.
For the second form Fermi made an analogy with electromagnetism, where the simplest multipole
operator M(Eλµ) corresponds to λ = 1 (remember that a λ = 0 photon does not exist). In this
0 0 0 0
dipolar √ √ that the matrix element is < l m |Y (E1µ)|lm >∝< l m |ρ|lm >, where ρ =
case one finds
(x − iy, 2z, x + iy)/ 2, i. e. it acquires a vector form. In beta decay the equivalent operator should be,
according to Fermi, the spin operator σ. This is related to the relativistic character of beta decay because
σ is an operator that appears naturally when writing the relativistic invariant version of the Schrödinger
equation (the Dirac equation). This second form, which is called Gamow-Teller, becomes,
Z
MGT = gA Ψ∗β(P ) (r)σΨα(N ) (r)dr (11.22)

where gA is the strength of the Gamow-Teller interaction.

The Pauli matrices σ = (σx , σy , σz ) only affect the spin degrees of freedom. Therefore σ does not
change the parity of the states involved in the transition. It does not change the orbital angular momentum
either, and lα = lβ . But the vector σ carries one unit of angular momentum. This implies that 0+ → 0+
transitions are forbidden and that the total final spin jβ is constraint by the triangular relation |jα − 1| ≤
jβ ≤ jα + 1.
The radial functions are also very similar to each other in this case, since they only differ by the spin
coupling (assuming nα =nβ ). Yet the matrix element (11.22) is smaller than the corresponding Fermi
matrix element because the recoupling of angular momenta induce a factor which is less than unity.
Comparisson with experiment has provided for the coupling constants the values gV ∼ 1.36 ×
10−3 e2 ~2 /(m2p c2 ), where mp is the proton mass, and gA /gV = −1.275. We have (to be checked)

gV = (1.41271 ± 0.00032) · 10−49 erg · cm3 , (11.23)

and
gA = (1.781 ± 0.011) · 10−49 erg · cm3 . (11.24)
At the quark level it is gV = −gA .

logaritm of f t

One sees from Eq. (11.19) that the bigger the matrix element M the smaller is the f t-value. That
implies that small values of f t corresponds to faster transition rates and shorter half lives. In the two
cases analyzed here, the Fermi transition is more likely than the Gamow-Teller one. Other choices of
the operator O may change parity and involve large differences in the transfer angular momenta. This
induces large differences between the corresponding radial functions and, as a result, the f t values can
change by many orders of magnitude. Due to this one defines the logaritm of f t. As seen in Table 11.1
the logf t indeed varies very much for various transitions. Besides the Fermi and Gamow-Teller, there
are other transitions with names which were introduced as more complex transitions were found.
This vast variations in the logf t values makes beta decay a powerful tool to determine the spin and
parities of nuclear states, as seen in the Exercises.
11 Beta and alpha readioactive decays 128

Table 11.1: Values of log10 f t and the corresponding restrictions upon the angular momentum transfer
∆J and parity change ∆π.
log10 f t ∆J lβ ∆π Decay type
2.9 - 3.7 0 0 no Supperallowed (Fermi)
3.8 - 6 0,1 0 no Allowed Gamow-Teller & Fermi
6 - 10 0,1,2 0,1 yes First forbidden
10 - 13 1,2,3 1,2 no Second forbidden
17-19 2,3,4 2,3 yes Third forbidden

Fermi decay
As we have seen before, the single-particle wave function contains a radial part and a spin-orbit part. The
radial part is determined by the principal quantum number n, the orbital angular momentum  l and
 the
total spin j. The parity of the state is (−1)l . In our case it is Ψα(N ) (r) = Rnα lα jα (r) Ylα (r̂)χ1/2 j m
α α
and the same for Ψ∗β(P ) (r). Integrating over the angles the matrix element becomes
Z
M = gV δlα lβ δjα jβ Rn∗ β lβ jβ (P ) (r)Rnα lα jα (N ) (r)r2 dr (11.25)

which shows that in the Fermi transition there is no change of parity or angular momentum. The residual
nucleus has the same spin and parity as the decaying nucleus.
The integral in Eq. (11.25) is very close to unity if nα =nβ since, except the Coulomb interaction, the
proton and neutron feel the same correlations. If isospin is conserved, and neglecting the influence of the
Coulomb field upon Rnβ lβ jβ (P ) (r), the neutron and proton radial wave functions are the same and
M = gV δnα nβ δlα lβ δjα jβ (11.26)
In this case the initial and final nuclear states are isobaric analogous, that is they differ only by their
isospin projections.
The operator associated with Fermi decay is proportional to the isospin raising and lowering operator.
As such it can only connect isobaric analogue states and it provides an exacting test of isospin conservation
in the nucleus. The operator for Fermi beta decay in terms of sums over the nucleons is
X
O(F± ) = tk± (11.27)
k

The matrix element is

2
B(F ) = hf |T± |ii (11.28)
where X
T± = t± (11.29)
k
is the total isospin raising and lowering operator for total isospin constructed out of the basic nucleon
isospin raising and lowering operators
t− |ni = |pi, t− |pi = 0 (11.30)
and
t+ |pi = |ni, t+ |ni = 0 (11.31)
The matrix elements obey the triangle conditions Jf = Ji (∆J = 0). The Fermi operator has πO = +1,
and thus the initial and final nuclear states must have πi πf = +1 for the matrix element to be nonzero
under the parity transform.
When isospin is conserved the Fermi matrix element must obey the isospin triangle condition Tf −
Ti (∆T = 0), and the Fermi operator can only connect isobaric analogue states. For β− decay
p
T− |ωi , Ji , Mi , Ti , Tzi i = Ti (Ti + 1) − Tzi (Tzi − 1)|ωi , Ji , Mi , Ti , Tzi − 1i (11.32)
and
2
B(F− ) = hωf , Jf , Mf , Tf , Tzi − 1|T− |ωi , Ji , Mi , Ti , Tzi i
  (11.33)
= Ti (Ti + 1) − Tzi (Tzi − 1) δωi ,ωf δJi ,Jf δMi ,Mf δTi ,Tf
11 Beta and alpha readioactive decays 129

For β+ we have
2
B(F+ ) = hωf , Jf , Mf , Tf , Tzi + 1|T+ |ωi , Ji , Mi , Ti , Tzi i
  (11.34)
= Ti (Ti + 1) − Tzi (Tzi + 1) δωi ,ωf δJi ,Jf δMi ,Mf δTi ,Tf

For neutron-rich nuclei (Ni > Zi ) we have Ti = Tzi and thus

B(F− )(Ni > Zi ) = 2Tzi = (Ni − Zi )δωi ,ωf δJi ,Jf δMi ,Mf δTi ,Tf (11.35)

and
B(F+ )(Ni > Zi ) = 0 (11.36)
For proton-rich nuclei (Zi > Ni ) we have Ti = −Tzi and thus

B(F+ )(Zi > Ni ) = −2Tzi = (Zi − Ni )δωi ,ωf δJi ,Jf δMi ,Mf δTi ,Tf (11.37)

and
B(F− )(Zi > Ni ) = 0 (11.38)

Gamow-Teller decay
The operator associated with Gamow-Teller decay also contains the nucleon spin operator. Since the
total spin S is not a good quantum number, Gamow-Teller beta decay goes in general to many final
states and provides a sensitive test of shell-model configuration mixing in the nucleus. The operator for
Gamow-Teller beta decay in terms of sums over the nucleons is
X
O(GT± ) = σk tk± (11.39)
k

The reduced matrix elements is

2 2
hf ||O(GT± )||ii Mi,f (GT± )
Bi,f (GT± ) = = (11.40)
2Ji + 1 2Ji + 1
The matrix elements are reduced in orbital space and the 2Ji + 1 factor comes from the average over
initial Mi states. The magnitude of reduced matrix element M (GT ) does not depend on the direction of
the transition, i.e.,
Mi,f (GT± ) = hf ||O(GT± )||ii (11.41)
The matrix elements obey the triangle ∆(Ji , ji , ∆J = 1). The Gamow-Teller operator has πO = +1 , and
thus the initial and final nuclear states must have πi πf = +1 for the matrix element to be nonzero under
the parity transform. When isospin is conserved the Gamow-Teller matrix elements obey the isospin
triangle condition ∆(Tf , Ti , ∆T = 1).
In second-quantized form the GT− operator has the form
X
O(GT− ) = hα|σt− |βia†α,p aβ,n (11.42)
αβ

where aβ,n destroys a neutron in state β and a†α,p creates a proton in state α. The J-coupled form is
 † λ
X aα,p ⊗ ãβ,n
O(GT− ) = hja , p||σt− ||jb , ni √ (11.43)
ja ,jb
2λ + 1

where λ = 1 for the GT operator. The reduced transition probability for the transition from an initial
state i to a final state f is given by
λ
hf || a†α,p ⊗ ãβ,n ||ii

X
B(GT− ) = hja , p||σt− ||jb , ni √ (11.44)
ja ,j
2λ + 1
b

The analogous equations for GT+ are

X
O(GT+ ) = hα|σt+ |βia†α,n aβ,p (11.45)
αβ
11 Beta and alpha readioactive decays 130

2hja ||s||jb i
Table 11.2: Gamow-Teller reduced single-particle matrix elements √
3
.
a/b s√
1/2 p3/2 p1/2 d5/2 d3/2 f7/2 f5/2
s1/2 2 √
p3/2 2 5/3 -4/3
√
p1/2 4/3 − 2/3 p √
d5/2 14/5
√ −4/√5
d3/2 4/ 5 −2/ 5 p p
f7/2 2p6/7 −4p 2/7
f5/2 4 2/7 − 10/7

where aβ,p destroys a proton in state β and a†α,n creates a neutron in state α. The J-coupled form is
 † λ
X aα,n ⊗ ãβ,p
O(GT+ ) = hja , n||σt+ ||jb , pi √ (11.46)
ja ,jb
2λ + 1

and the reduced transition probability is

λ
hf || a†α,n ⊗ ãβ,p ||ii

X
B(GT+ ) = hja , n||σt+ ||jb , pi √ . (11.47)
ja ,jb
2λ + 1

The reduced single-particle matrix elements are given by

hja , p||σt− ||jb , ni = hja , n||σt+ ||jb , pi = 2hja ||s||jb i (11.48)

where the matrix elements of s are given by

 
la +ja +3/2
p 1/2 1/2 1
hja ||s||jb i =(−1) (2ja + 1)(2jb + 1) hs||s||siδla ,lb δna ,nb (11.49)
jb ja la

with p
hs||s||si = 3/2. (11.50)
The orbits which are connect by the GT operator are very selective since the matrix elements of s has
the selection rules δla ,lb and δna ,nb .
We have
hja ||s||jb i = (−1)ja +jb +1 hjb ||s||ja i = (−1)δab +1 hjb ||s||ja i. (11.51)
In practice, we have hj> ||s||jb i > 0 and hj< ||s||jb i < 0 where j>(<) = l + 1/2(|l − 1/2|). This is because
the 6j symbols in above equation take positive values for ja = jb and negative values for ja 6= jb .

Effective operators for the Gamow-Teller decay

There are several reasons why the “free-nucleon” calculations may differ from experiment. In reality
the nuclear wave functions are more complicated than the theoretical model we use in that they may
incorporate nucleon degrees of freedom beyond the shell space. In addition, non-nucleonic degrees of
freedom involving the delta isobars and mesons in the nucleus may also be important. The calculation of
the corrections corresponding to these processes have been the subject of many theoretical investigations.
2
Since the factor gA appears in front of B(GT ) one might parameterize the effective matrix in terms
of an effective gA value, as we are doing in electromagnetic transition studies. In nuclei we have gA ∼ 1.

11.3 Sum rules

Sum rules for Fermi and Gamow-Teller matrix elements can be obtained. The sum rule for Fermi is
obtained from the sum
X  Xh 2 2
i
Bi,f (F− ) − Bi,f (F+ ) = hf |T− |ii − hf |T+ |ii (11.52)
f f
11 Beta and alpha readioactive decays 131

The final states f in the T− matrix element go with the Zf = Zi + 1 nucleus and those in the T+ matrix
element to with the Zf = Zi − 1 nucleus. One can explicitly sum over the final states to obtain
X 
hi|T+ |f ihf |T− |ii − hi|T− |f ihf |T+ |ii = hi|T+ T− − T− T+ |ii = hi|2Tz |ii = Ni − Zi (11.53)
f

The sum rule for Gamow-Teller is obtained as follows.

X X 2 X X 2
hf | σk,µ tk− |ii − hf | σk,µ tk+ |ii
f,µ k f,µ k
X X X X X X
= hi| σk,µ tk+ |f ihf | σk0 ,µ tk0 − |ii − hi| σk,µ tk− |f ihf | σk0 ,µ tk0 + |ii
f,µ k k0 f,µ k k0
" ! ! ! ! #
X X X X X (11.54)
= hi| σk,µ tk+ σk0 ,µ tk0 − − σk,µ tk− σk0 ,µ tk0 + |ii
µ k k0 k k0
X X X
2
= hi| σk,µ [tk+ tk− − tk− tk+ ]|ii = 3hi| [tk+ tk− − tk− tk+ ]|ii
µ k k

=3hi|T+ T− − T− T+ |ii = 3hi|2Tz |ii = 3(Ni − Zi )

We have used the fact that σx2 = σy2 = σz2 = 1 When k 6= k 0 the operators commute and cancel. These
sum rules hold for each Mi value and thus also hold for B(F ) and B(GT ) when we take an average over
the Mi values. Thus X 
Bi,f (F− ) − Bi,f (F+ ) = Ni − Zi (11.55)
f

and X 
Bi,f (GT− ) − Bi,f (GT+ ) = 3(Ni − Zi ) (11.56)
f

The sum-rule for the Fermi matrix elements applies even when isospin is not conserved. For and N > Z
we usually have Ti = Tzi which means that B(F+ ) = 0 and we can obtain B(F− ) = N i − Zi for the
transition to the isobaric analogue state. For N = Z(Tzi = 0) and Ti = 0 we have B(F+ ) = B(F− ) = 0,
and for Ti = 1 we have B(F+ ) = B(F− ) = 2. Fermi transitions which would be zero if isospin is conserved
are called isospin-forbidden Fermi transitions.
When N > Z there are some situations where one has B(GT+ ) = 0, and we have B(GT− ) =
3(N i − Zi). In particular for the β− decay of the neutron we have B(F− ) = 1 and B(GT− ) = 3.

14
11.4 C-dating beta decay
The radioisotope 14 C is the basis for radiocarbon dating in which it is assumed that living organisms have
a 14 C/12 C ratio which is the same as that in the atmosphere and that this atmospheric ratio has been
constant in the past. The interaction between the organism and the atmosphere stops at death, and the
14
C that remains in the organism decays. By measuring the 14 C/12 C ratio directly (using Accelerator
Mass Spectrometry) or the rate of decay of 14 C (using radiometric methods), the time since death can be
estimated. Clearly there is another assumption underlying this process, namely that the decay of 14 C is
exactly exponential for all time. The half-life for 14 C was initially found to be 5, 568 ± 30 yr (the Libby
half-life) but was later changed to 5, 730 ± 40 yr (the Cambridge half life).
The anomalously long β decay half-life of 14 C has been of continuous theoretical interest since the
appearance of the nuclear shell model [3, 4, 5, 6]. The decay involves the J π = 0+ ground state of 14 C
and the J π = 1+ ground state of 14 N and satisfies the selection rule for typical allowed Gamow-Teller
(GT) transitions. However, the extracted transition amplitude from experimental half-life is thousands
of times smaller than that of allowed transitions. The inhibition should be attributed to the accidental
cancellation of certain components of the wave functions of the involved states that contribute to the
transition. It was recognized long ago that the tensor part of the nuclear force play an essential role in
inducing the cancellation [7, 8].
In the original paper of Jancovici and Talmi’s [7], an unreasonably large tensor force was introduced
to induce the cancellation. Later studies show that this problem can be rectified by redefining the
radial dependence of the tensor component (for reviews, see Ref. [4]). One may expect that the shape
11 Beta and alpha readioactive decays 132

and strength of the tensor force were confined in realistic nucleon-nucleon (NN) potentials which are
determined by fitting NN scattering observables. However, the studies of Zamick and collaborators [5, 9,
10] showed that the cancellation cannot be reproduced by calculations with realistic Hamiltonians [11]
derived from microscopic NN potentials like the Hamada-Johnston potential or the Bonn potential. This
failure was also seen in very recent calculations of Refs. [6, 12, 13, 14] with modern one-boson-exchange
and chiral potentials. It was suggested that the problem may indicate that the tensor component of
the in-medium NN interaction is much weaker than that of the bare potential [5, 9, 10]. Holt et al.
claimed that the problem of realistic calculations in reproducing the long β decay half-life of the 14 C
can be solved by taking into account the Brown-Rho scaling in-medium modification [6] or three-nucleon
corrections [12] of the N N interaction. No-core shell model calculations by Maris et al. [14] also show
that the cancellation can be reproduced by introducing a three-nucleon force of chiral perturbation theory.
The wave functions of the 14 C (J π = 0+ , T = 1) and 14 N (J π = 1+ , T = 0) ground states, denoted
by |ψi i and |ψf i, respectively, can be well described as two holes occupying the 0p1/2 and 0p3/2 single-
particle orbits by assuming 16 O as the inert core [7, 8]. (Recent large space calculations of Ref. [14]
also suggested that the cancellation mostly occurs in the p shell.) In the jj coupling scheme, the wave
functions can be written as [9]

|ψi i = κ|0p−2 −2
1/2 i + η|0p3/2 i,

|ψf i = a|0p−2 −1 −1 −2
1/2 i + b|0p3/2 0p1/2 i + c|0p3/2 i, (11.57)

where the κ and η and a, b and c denote the corresponding wave function amplitudes. The GT transition
matrix element is determined by
√ √ i
r h
2
M (GT) = hψf ||στ ||ψi i = κ(a + 2b) + η( 2b − 5c) . (11.58)
3
In many cases the wave functions of 14 C and 14 N were calculated in the LS coupling scheme [7]. The
transformation between wave functions in LS and jj coupling schemes is known in analytic forms in
terms of 6j and 9j symbols. To facilitate the comparison between wave functions in different coupling
schemes available on the market, the explicit expressions for the transformation are listed below as
√  !
|0p−2 i
 1  
| S0 i 1 1 2
=√ √ 1/2
, (11.59)
|3 P0 i 3 2 −1 |0p−2
3/2 i

and √
|0p−2
 
|3 S1 i i
  
−1
√ −4
√ √ 10 1/2
1
 | 1 P1 i  = √  −1 −1
6 √6 15   |0p3/2 0p1/2 i  . (11.60)
 
3 √ √
|3 D1 i 20 − 5 − 2 |0p−2
3/2 i

In evaluating above wave function amplitudes and the transition amplitude M (GT), one may start
from empirical as well as realistic shell-model interactions. It has been established that realistic inter-
actions derived from bare NN potentials are in general close to empirical ones which are obtained by
fitting experimental data [15, 16, 17, 18]. This can also be seen from Table 11.3 where the diagonal ma-
trix elements of some well-defined empirical p-shell interactions [19, 20, 21, 22] and those of the realistic
interactions are listed for comparison.
The realistic interactions in Table 11.3 are derived from the underlying N N potentials for which we
take the state-of-the-art Bonn potential (CD-Bonn) of Ref. [23] and the chiral potential (N3 LO) of Ref.
[24]. The have to be renormalized to include contributions from excluded configurations and to avoid the
hard core of the bare N N potential. In this work the short-range repulsion is taken into account with the
momentum-space renormalization group decimation method (Vlow−k ) [25] and the G-matrix approach
[15]. In the Vlow−k method, high momentum contributions are integrated out with the introduction
of a cutoff in momentum space for which we take Λ = 2 fm−1 . The G-matrix is calculated by using
the double-partitioned approach with the standard starting energies of ω = −5, -20, -50, -90 and -140
MeV [15] and the HO parameter of ~ω0 = 45A−1/3 − 25A−2/3 ≈ 14.37 MeV. The detailed description
of the renormalized G-matrix calculation can be found in Ref. [15] with a public code available. Core
polarization effects upto the second order in perturbation theory are taken into account through the
folded-diagram method. The projection operator appearing in the calculation is defined with the HO
orbital boundaries (see Fig. 31 of Ref. [15] as an illustration) of n1 =1, n2 =3 and n3 =36. It means we
take 4 He as the closed core, 0p3/2 and 0p1/2 as the valence single-particle orbits and we take into account
11 Beta and alpha readioactive decays 133

Table 11.3: Comparison between diagonal matrix elements hij|V |ijiJT (in MeV) of empirical and realistic
interactions. The realistic interactions are calculated from the CD-Bonn and N3 LO potentials with the
G-matrix (G) and Vlow−k (K) renormalization approaches. Only the few terms that related to the
description of the dating β decay are listed for simplicity.

J = 0, T = 1 J = 1, T = 0
Interaction 0p23/2 0p21/2 0p23/2 0p3/2 0p1/2 0p21/2
Empirical
CK [19] -3.19 -0.26 -3.58 -6.22 -4.15
VWG [20] -3.68 -0.15 -2.62 -6.55 -3.95
WBT [21] -3.85 -1.22 -4.16 -6.86 -3.45
WBP [21] -3.91 -1.15 -3.86 -6.94 -3.45
Realistic
CD-Bonn(K) -4.26 -0.77 -3.30 -8.26 -3.56
N3 LO(K) -4.06 -0.66 -3.23 -8.18 -3.65
CD-Bonn(G) -4.09 -0.74 -3.24 -8.20 -3.69
N3 LO(G) -3.86 -0.60 -3.86 -8.10 -3.74

the influence from higher-lying HO orbitals with quantum number N = 2n + l ≤ 7. The calculation is
done in the isospin space. The charge independence breaking effect of the N N potential is neglected for
simplicity.
The empirical interactions of Refs. [19, 20, 21] are constructed in the particle-particle channel by
assuming 4 He as the inert core and the single-particle energies as free parameters. In above cases it was
assumed that the interactions are are same in the hole-hole channel. The energy splitting between 0p−1 1/2
and 0p−13/2 orbits in 15
C and 15
N is calculated to be ε = ε(0p−1
3/2 ) − ε(0p−1
1/2 ) = 6.3 MeV [19], 7.3 MeV [20]
and 6.5 MeV [21]. The calculations are very close to the experimental result, i.e, ε = 6.3 MeV.
The ground state of 14 C is dominated by the configuration of |0p−2 1/2 i due to the large spin-orbit
−1 −1
splitting between orbits 0p1/2 and 0p3/2 . This is supported by calculations with both empirical and
realistic interactions, as seen from Table 11.4. As a result, the coefficient κ of Eq. (1) is significantly larger
than η (κ and η have the same sign and κ = 1 and η = 0 in the single-particle limit). The mixing of the two
corresponding configurations is induced by the non-diagonal matrix element h0p23/2 |V |0p21/2 iJ=0,T =1 for
which realistic and empirical interactions give a similar strength. Comparison between the non-diagonal
matrix elements is done in Table 11.5.
11 Beta and alpha readioactive decays 134

Table 11.4: Comparison between different wave functions calculated with empirical and realistic interac-
tions. All calculations are done with the code [26] except those of Jancovici and Talmi’s and of the chiral
potential which are taken from Ref. [7] and Ref. [13], respectively.

Interaction η κ c b a
Empirical
CK [19] 0.38 0.92 -0.027 -0.31 0.95
VWG [20] 0.36 0.93 -0.063 -0.27 0.96
WBT [21] 0.31 0.95 0.033 -0.43 0.90
WBP [21] 0.30 0.95 0.014 -0.41 0.91
JT [7] 0.09 0.99 0.20 -0.41 0.89
Zamick [9] 0.22 0.98 0.014 -0.40 0.92
VF [27] 0.25 0.97 0.12 -0.36 0.97
Realistic
CD-Bonn(K) 0.40 0.92 0.20 -0.77 0.61
N3 LO(K) 0.39 0.92 0.15 -0.71 0.69
CD-Bonn(G) 0.39 0.92 0.14 -0.70 0.70
N3 LO(G) 0.38 0.93 0.11 -0.65 0.75
Chiral [13] 0.40 0.92 0.14 -0.68 0.72
 11

Table 11.5: Same as Table 11.3 but for the non-diagonal matrix elements hij|V |kliJT (in MeV) of empirical and realistic interactions.

ijklJT CK [19] VWG [20] WBT [21] WBP [21] CD-Bonn(K) N3 LO(K) CD-Bonn(G) N3 LO(G)
113301 -4.86 -4.99 -3.84 -3.70 -4.92 -4.77 -4.77 -4.58
111310 1.69 1.71 1.81 1.70 1.82 1.99 1.96 2.09
Beta and alpha readioactive decays

Table 11.6: The central (C), spin-orbit (SO) and tensor (T) components of the matrix elements hij|V |kliJT of empirical and realistic interactions.

VWG WBT CD-Bonn(K) N3 LO(G)

ijkl C SO T C SO T C SO T C SO T
J π = 0+ , T = 1
1111 -1.66 0.58 0.92 -1.33 054 -0.42 -2.14 0.37 1.00 -2.02 0.38 1.04
3333 -4.43 0.29 0.46 -3.96 0.33 -0.21 -4.93 0.19 0.50 -4.56 0.17 0.52
1133 -3.93 -0.41 -0.65 -3.72 -0.42 0.29 -3.94 -0.27 -0.71 -3.58 -0.26 -0.73
J π = 1+ , T = 0
1111 -4.27 0.41 -0.098 -4.49 1.11 -0.075 -4.64 -0.10 1.17 -4.65 -0.15 1.15
1113 1.08 -0.21 0.83 0.43 -0.0019 1.38 1.17 0.26 0.39 1.29 0.30 0.50
1313 -5.85 0.10 -0.80 -5.85 0.35 -1.36 -7.70 0.13 -0.69 -7.46 0.15 -0.78
1133 2.72 -0.13 -0.46 1.55 -0.046 -0.82 3.80 0.07 -0.99 3.76 0.18 -1.03
1333 3.72 0.065 -0.016 2.41 0.045 -0.012 5.73 0.01 0.19 5.49 0.02 0.18
3333 -2.96 0.041 0.30 -4.11 -0.58 0.53 -3.49 -0.32 0.51 -3.34 -0.36 0.54
135
11 Beta and alpha readioactive decays 136

Similarly, one may safely expect that |0p−2 1/2 i should be the dominated configuration in the ground
14
state wave function of N since the other two configurations lie at much higher energies. This expectation
is supported by all calculations with empirical interactions listed in Table 11.4. Since the amplitude κ is
much larger than η, the suppression of the Gamow-Teller transition strength [Eq. (2)] should be largely
due to the cancellation between a and 2b (a and b have different signs). That is, the term |0p−1 −1
3/2 0p1/2 i
should most likely be the second largest component in the ground state wave function of 14 N. Most of our
calculations with different interactions predict small values for the absolute value √ of the wave function
amplitude of configuration |0p−2 3/2 i. In this case we should have a ∼ −(2κ + 2η)b in reproducing the
cancellation.
It can be seen from Table 11.4 that the problem of the realistic calculation in reproducing the can-
cellation is related to the fact that the predicted amplitude a (b) is significantly smaller (larger) than
expected. The ratio between the amplitudes a and b are sensitive to the strengths of diagonal matrix
elements h0p21/2 |V |0p21/2 iJ=1,T =0 and h0p3/2 0p1/2 |V |0p3/2 0p1/2 iJ=1,T =0 and the non-diagonal matrix ele-
ment h0p3/2 0p1/2 |V | 0p21/2 i J=1,T =0 . If we neglect the contribution from the configuration |0p23/2 i and
restrict the calculation to dimension two, the mixing between above two components would be solely
dominated by the perturbation term
h0p3/2 0p1/2 |V |0p21/2 i
. (11.61)
ε + h0p3/2 0p1/2 |V |0p3/2 0p1/2 i − h0p21/2 |V |0p21/2 i

By comparing the interaction matrix elements listed in Tables 11.3 & 11.5, it can be seen that the problem
of realistic interaction is that the strength for diagonal matrix element h0p3/2 0p1/2 |V |0p3/2 0p1/2 iJ=1,T =0
it predicted is much larger than empirical ones.
It may still be of interest to figure out the way how the tensor force induces the cancellation in the
jj coupling scheme (The impossibility of inducing cancellation with effective interactions without tensor
force was first shown by Inglis [3]). Although the different components are mixed in usual construction
of empirical interactions, it is possible to separate the central, spin-orbit (vector) and tensor force com-
ponents of the effective interaction through the spin-tensor decomposition procedure [28]. As examples
in Table 11.6 is listed the different components of realistic interactions and the empirical interactions of
Refs. [20, 21]. The decompositions of various Cohen-Kurath interactions can be found in Ref. [29] and will
not be presented here for simplicity. It can be seen from the figure that for the diagonal matrix element of
concern, h0p3/2 0p1/2 |V |0p3/2 0p1/2 iJ=1,T =0 , empirical and realistic interactions predict similar strength
for the spin-orbit and tensor component. Large difference is only seen in the central channel. But as
pointed out in Ref. [28], there is no trivial relation between the spin-tensor decomposition of the effective
interaction and different components of the underlying N N potential. Especially, the tensor force of the
N N potential may contribute significantly to the overall effective interaction when renormalization and
core-polarization effects are taken into account [11].
In Table 11.7 is listed the wave functions calculated with the tensor force component removed from
the effective interactions. In this case it is seen that the ground state wave function of 14 N is over-
whelmingly dominated by the configuration of |0p−2 1/2 i. This is because the non-diagonal matrix element
2 J=1,T =0
h0p3/2 0p1/2 |V |0p1/2 i , which is crucial in inducing the configuration mixing, contain important
contribution from the tensor force.

11.5 Alpha decay

In this form of decay the nucleus emits a 4 He2 nucleus, i. e. an alpha particle. This is a traditional
decay process which, as already mentioned above, was discovered even before beta and gamma decay.
The physical mechanism which induces this form of decay has been found to be also responsible for other
similar processes, like cluster and proton decay. We will study all these processes together, since formally
there is no diference among them. However, we will concentrate in the alpha particle first because that
was the historical sequence.
Alpha decay is the emission of an alpha particle from an unstable state in the decaying nucleus, i. e.

A→B+α (11.62)

where A is called ”mother mucleus”, B ”daughter mucleus” and α is the outgoing alpha particle.
11 Beta and alpha readioactive decays 137

Table 11.7: Wave functions of 14 C and 14 N calculated with the central force and central and spin-orbit
force components of empirical and realistic effective interactions.

Interaction η κ c b a
Central force only
CK [19] 0.33 0.94 -0.15 -0.085 0.99
VWG [20] 0.29 0.96 -0.14 -0.086 0.99
WBT [21] 0.30 0.95 -0.11 -0.033 0.99
CD-Bonn(K) 0.33 0.94 -0.30 0.14 0.94
N3 LO(K) 0.32 0.95 -0.20 -0.017 0.98
CD-Bonn(G) 0.32 0.95 -0.27 0.0080 0.96
N3 LO(G) 0.30 0.95 -0.25 0.037 0.97
Central plus spin-orbit
CK [19] 0.34 0.94 -0.15 -0.085 0.99
VWG [20] 0.32 0.95 -0.15 -0.053 0.99
WBT [21] 0.33 0.94 -0.12 -0.030 0.99
CD-Bonn(K) 0.35 0.94 -0.27 0.04 0.96
N3 LO(K) 0.34 0.94 -0.18 -0.11 0.98
CD-Bonn(G) 0.34 0.94 -0.25 -0.0071 0.97
N3 LO(G) 0.32 0.95 -0.23 -0.054 0.97

The theoretical study of this process was performed by Gamow in 1928, that is at the beginning of
quantum mechanics. Gamow’s explanation of the decay was the first application of the probabilistic
interpretation of quantum mechanics.
Gamow assumed that the alpha particle exists on the surface of the mother nucleus at an excitation
energy which is above the continuum threshold. From here it escapes the nucleus with a kinetic energy
which, at large distances, takes the value E. This cannot happen in classical mechanics, since at that
energy the kinetic energy of the particle would be imaginary while penetrating the barrier, that is the
region denoted by a dashed line in Fig. (11.1). But Gamow realized that in quantum mechanics the
centrifugal and Coulob barriers would trap the alpha particle inside the nucleus during a time before
decaying. He evaluated the probability of the particle penetrating the barrier seen in Fig. 11.1 and
could thus explain the relation between the decay half life and the kinetic energy of the escaping alpha
particle. This was a tremendous success not only for Gamow himself but also, and specially, for quantum
mechanics as a subject, since due to its apparent contradictions quantum mechanics was questioned by
outstanding physicists at that time.
To evaluate the alpha decay half life we will consider the motion of the outgoing alpha particle. Inside
the nucleus the two neutrons and two protons which eventually will form the alpha particle move under
the influence of the interactions of the other nucleons. A proper microscopic calculation should be able
to describe this motion and also the mechanism that clusters those neutrons and protons in the nuclear
surface to become the alpha particle. We will not go into that mechanism, but rather assume that the
alpha particle is already formed just inside the nuclear surface. The 4 He2 nucleus thus formed will be
considered a spinless point particle located at a distance r from the center of the mother nucleus. We
(int)
will denote the corresponding wave function by Ψlm (r), where l is the alpha particle orbital angular
momentum and m its z-projection. The label ”(int)” indicates that this is the solution of the Schrödinger
equation inside the nucleus. Since the intrinsic spin of the alpha particle is zero l is also the total angular
momentum. When the alpha particle leaves the nucleus the only interaction it feels is the Coulomb force
from the daugther nucleus. The solution of the corresponding Schrödinger equation, which also includes
the centrifugal term, is known analytically. It has the form

(out) Gl (r) + iFl (r)

Ψlm (r) = Nl Ylm (r̂) (11.63)
r
Where Fl (Gl ) is a real function called the ”regular” (”irregular”) Coulomb function and Ylm (r̂) is the
spherical harmonics. Since we assume spherical symmetry the angular momentum projection m does
not influence the physics of the problem. We include it in the wave functions, e. g. in Eq. (11.63), for
mathematical completeness.
11 Beta and alpha readioactive decays 138

V(r)

0 R r

Figure 11.1: Unstable state at energy E from where the alpha particle is emitted. The penetrability is
determined by the width of the barrier at the energy of the state, indicated by a dashed line.

At very large distances, where the Coulomb and centrifugal forces vanish, it is

Gl (r) + iFl (r) −→ eikr (11.64)

r→∞

where k is the wave number related to the kinetic energy E of the alpha particle by

~2 k 2
E= (11.65)
2µ
µ is the alpha particle reduced mass. We can now evaluate the alpha decay rate, that is the probability rate
per second that the alpha particle is emitted. For this we first notice that the measurement of the decaying
alpha particle is performed at very large distance in comparisson to the nuclear size, since the nuclear size
is of the order of 10−11 cm while the distance where the detector of the alpha particle is located is of the
order of 102 cm. At that distance the alpha particle moves following a radial trajectory, i. e. its velocity
is v = vr/r, where v = ~k/µ. The number of alpha particles going through a volume dV = dSdr, where
(out)
dS is a surface perpendicular to v, is, at large distance, dNl = |Ψlm (r)|2 dV = |Nl eikr Ylm (r̂)|2 dV /r2 .
2
In spherical coordinates it is dS = r sinθdθdφ. Therefore the number of particles per unit time going
through the surface dS is
dNl dr
= |Nl |2 |Ylm (θφ)|2 sinθdθdφ (11.66)
dt dt
Integrating over the angles one gets |Ylm |2 (θφ)sinθdθdφ=1 and since dr/dt = v the mean life τ , which
R

is defined by 1/τ = dNl /dt, is

1
= |Nl |2 v (11.67)
τ
The width Γl of the resonance is related to the mean life by the energy-time relation Γl τ = ~.
The constant Nl is obtained by matching the internal and external wave functions just outside at the
surface of the nucleus, where the nuclear interaction is negligible and only the Coulomb and centrifugal
forces are active. Calling that radius R one gets,
(int)
RΨl (R)
Nl = (11.68)
Gl (R) + iFl (R)
11 Beta and alpha readioactive decays 139

Thomas expression for the width

Collecting all factors together we finally get for the width the expression,
(int)
~2 k R2 |Ψl (R)|2
Γl = (11.69)
µ Fl2 + G2l

The penetrability through the Coulomb and centrifugal barriers is

kR
Pl (R) = (11.70)
Fl2 + G2l

In terms of the penetrability the width acquires the form,

(int)
~2 kR|Ψl (R)|2
Γl = Pl (R) (11.71)
µ
The penetrability is an exponential function of the energy and is the determinant factor in the value of
the width. If the energy increases, then the penetrability also increases, as can be seen in Fig. (11.1).
Therefore the probability that the particle is emitted increases strongly with increasing energy. That is,
the half life decreases and the width increases.
The centrifugal barrier increases with the angular momentum. This implies that an alpha unstable
nucleus will decay to a state that corresponds to the largest transition energy and the minimun angular
momentum transfer.
The Gamow expression for the width contains the penetrability Pl but for the other factors in Eq.
(11.71) he used a phenomenological form and instead of those factors he included a free parameter,
which could be interpreted as a hitting frequency of the alpha particle on the walls of the nucleus. The
expression (11.71) was obtained much later.
In principle one can evaluate the alpha decay width microscopically by using the shell model. One
thus determines the probability that the alpha particle is formed on the nuclear surface and also whether
the four nucleons that constitute the alpha particle are indeed clustered in a point. One then matches
this function with the external function (11.63), as we did above. The width should not depend upon
the matching point R. This is an important point, since if Γl (R) depends on R then that is a sign that
the calculation is defective.
In the late 1970’s and early 1980’s other radioactive decays were discovered, where instead of alpha
particles heavier nuclei were emitted. These processes are called ”cluster decay”. Also proton decay was
discovered, in which the emitted particle is a proton. All these cases can be evaluated in the same fashion
(int)
as we did above. The only difference is that the internal wave function Ψlm (r) should correspond to
the case under study. This is rather straightforward to do for the proton case, but for heavy clusters it
is a formidable task which was barely attempted so far.
11 Beta and alpha readioactive decays 140

11.6 Homework problems

Exercise 1:
a) Read the literature and describe briefly the reason why the neutrino was introduced as an elemen-
tary particle and which law requires that in the neutron beta decay an antineutrino is emitted?
c) Explain why the beta decay p → n + e+ + ν̄ and the electron capture p + e− → n + ν̄ are not
possible.
− −
d) Why the beta decay 147 147 147 147
68 Er → e + ν̄ + 69 T m is possible but 68 Er → e + ν̄ + 63 Eu is not?.

Exercise 2:
a) Determine in the decay scheme shown in Fig. 11.2 whether it is β + or β − the decay from the state
3 and from the state 3− .
+

b) Determine the most likely spins and parities of the states in the nucleus (N − 1, Z + 1) and the
electromagnetic transitions among the levels in that Figure.

3-
3+ (N-2,Z+2)
(N,Z)
logft=6 logft=2.8

logft=4 logft=7

0+
(N-1,Z+1)
Figure 11.2: Beta decays of the odd-odd nuclei (N, Z) and (N − 2, Z + 2).

c) Assign the most likely spins and parities, and the corresponding electromagnetic transition in the
nucleus (N − 1, Z + 1) of Fig. 11.3

10-
(N,Z) logft=20

logft=19

(N-1,Z+1)
Figure 11.3: Beta decay of the odd-odd nucleus (N, Z).

Exercise 3:
Determine the order of importance regarding the transition probabilities among the alpha decay
transitions shown in Fig. 11.4.

0+
(N+2,Z+2) (c)
(b)
2+

2+
(a)

0+
(N,Z)
Figure 11.4: Alpha decay scheme of the even-even nucleus (N + 2, Z + 2).
11 Beta and alpha readioactive decays 141

Exercise 4:
a)Give the coupling constants of beta decay (GV and GA ) and compare them with those of EM,
Strong interaction and gravitation. Is the weak interaction weaker than gravitation?
b) How does an experimental nuclear physicist proceed to determine the log f t value in a beta decay
transition?

Exercise 5:
Calculate the Fermi and Gamow-Teller partial decay half-lives of the neutron. How long is the total
half-life?
 Bibliography

[1] Colloquium: The neutron lifetime Fred E. Wietfeldt and Geoffrey L. Greene Rev. Mod. Phys. 83,
1173 (2011).
[2] J. Liu et al. (UCNA Collaboration) Phys. Rev. Lett. 105, 181803 (2010). Determination of the
Axial-Vector Weak Coupling Constant with Ultracold Neutrons

[3] D.R. Inglis, Rev. Mod. Phys. 25 390 (1953).

[4] I. Talmi, Fifty Years of the Shell Model−The Quest for the Effective Interaction, Adv. Nucl. Phys.
27 1 (2003).
[5] M.S. Fayache, L. Zamick, and H. Müther, Phys. Rev. C 60 067305 (1999).

[6] J.W. Holt, G.E. Brown, T.T.S. Kuo, J.D. Holt, and R. Machleidt, Phys. Rev. Lett. 100 062501
(2008).
[7] B. Jancovici and I. Talmi, Phys. Rev. 95 289 (1954).
[8] H. J. Rose, O. Häusser, and E. K. Warburton, Rev. Mod. Phys. 40 591 (1968).

[9] L. Zamick, Phys. Lett. 21 194 (1966).

[10] D.C. Zheng and L. Zamick, Ann. Phys. (N.Y.) 206 106 (1991).
[11] T.T.S. Kuo, G.E. Brown, Phys. Lett. 18 54 (1965).

[12] J.W. Holt, N. Kaiser and W. Weise, Phys. Rev. C 79 054331 (2009).
[13] J.W. Holt, N. Kaiser and W. Weise, arXiv: 1011.6623.
[14] P. Maris, J.P. Vary, P. Navratil, W.E. Ormand, H. Nam, and D.J. Dean, Phys. Rev. Lett. 106
202502 (2011).

[15] M. Hjorth-Jensen, T.T.S. Kuo, and E. Osnes, Phys. Rep. 261 125 (1995).
[16] A. Poves and A.P. Zuker, Phys. Rep. 70 235 (1981).
[17] B. A. Brown and W. A. Richter Phys. Rev. C 74 034315 (2006).
[18] M. Honma, T. Otsuka, T. Mizusaki, and M. Hjorth-Jensen, Phys. Rev. C 80 064323 (2009).

[19] S. Cohen and D. Kurath, Nucl. Phys. 101 1 (1967).

[20] A. G. M. van Hees, A. A. Wolters, P. W. M. Glaudemans, Nucl. Phys. 476 61 (1988).
[21] E.K. Warburton and B.A. Brown, Phys. Rev. C 46 923 (1992).

[22] W. T. Chou, E. K. Warburton, and B. A. Brown, Phys. Rev. C 47 163 (1993).

[23] R. Machleidt, Phys. Rev. C 63 024001 (2001).
[24] D.R. Entem and R. Machleidt, Phys. Rev. C 68 041001(R) (2003).
[25] S.K. Bogner, T.T.S. Kuo, and A. Schwenk, Phys. Rep. 386 1 (2003).

[26] C. Qi and F.R. Xu, Chin. Phys. C 32 (S2) 112 (2008).

[27] W. Visscher and R. Ferrell, Phys. Rev. 107 781 (1957).
[28] M.W. Kirson, Phys. Lett. B47 110 (1973).

[29] K. Yoro, Nucl. Phys. A333 67 (1980).

Chapter 12
Nuclear energy and Nucleosynthesis

Fission. Energy production. Chain reactions. Fusion. Nucleosynthesis in stars. Hydrogen burning: the
pp chain and CNO cycle. Helium burning. The burning of heavy elements. The end of the star.

The subject of this Chapter is vast and it is beyond the scope of this course to go into details of the
calculations that will be presented here. Our aim is to give the main concepts leading to the explanation
of the dynamic fission and fusion processes and the appearence of the observed matter in the Universe.

12.1 Fission

We have seen at the end of last Chapter that one of the channels taken by unstable nuclei in their way to
stable configurations is alpha decay. In this case, as well as in all radioactive cluster decays, the nucleons
that eventually constitute the emitted particle get clustered before emission. The formation of the cluster
and its penetration through the Coulomb and centrifugal barriers takes a certain time which in average
is the decay half life. The emitted particle escapes the daughter nucleus with an energy which is the
Q-value of the process. It corresponds to the relative kinetic energy between the daughter nucleus and
the particle. Since we will use this concept extensively in this Chapter, it is worthwhile to describe it
shortly once again.
In the radioactive decay process
M →D+C (12.1)
where M is the mother nucleus, D the daughter and C the emitted cluster, energy conservation implies
that the Q-value is
Q = B(D) + B(C) − B(M ) (12.2)
If the Q-value is positive then the decay occurs spontaneously. But there is another very similar process
in which the mother nucleus breaks up spontaneously into two or more pieces, each piece consisting of a
lighter nucleus. This is called fission.
Fission occurs spontaneously only in heavy nuclei. This is because the binding energy per nucleon
increases up to the isotope F e (Z=26) and for heavier isotopes it decreases steadily. Therefore light
isotopes cannot fission, since the smaller binding energies of the outgoing particles with respect to the
mother binding energy makes that the Q-value becomes negative.
Heavy isotopes usually decay by emitting alpha particles. This is because the binding energy of
the alpha particle is large, and the nucleus decay by alpha emission before the fission process starts.
Therefore it is not spontaneous fission which makes this process so important in nuclear physics. Rather
it is induced fission which has become a very important tool in nuclear physics research and applications.
The most common and useful induced fission is by bombarding a heavy isotope with neutrons at very
small energy (that is moving slowly). In this condition the neutron will be absorbed by the target and
the mother nucleus thus formed becomes unstable triggering the fission process described above. The
neutron should carry an energy of about 0.025 eV only. These low energy neutrons are called ”thermal
neutrons”.
This process is used in experimental nuclear physics to reach isotopes which are otherwise very difficult
to detect. An example is the measurement of the spectrum of 141 Ba (Barium 141). This nucleus is
produced in the reaction
235
U + n →236 U −→ 141 Ba +92 Kr + 3n (12.3)
f ission

in which also are produced Krypton 92 and three neutrons, with the release of huge amounts of energy.
With the help of magnetic fields the nucleus 141 Ba is directed to the measurement chamber.
The energy released in this reaction is used in energy production applications, as we will see below.
12 Nuclear energy and Nucleosynthesis 144

Energy production

The big fragments that result of the partition of the nucleus are isotopes with a definite proton value Z.
But the number N of neutrons may differ by a few units. Therefore in the fission of a nucleus there is
not an exact number of neutrons which are emitted, In particular the neutrons emitted in the reaction
(12.3) are three in average. The energies of these neutrons are of the order of 1 MeV and, therefore, are
many orders of magnitude larger than what is needed for any of them to start a new fission reaction.
This is what happened in the first attempt, in the 1930’s, to extract energy from the fission process. It
was Fermi who realized that if one mixes the fissioning nuclei with a material that scatters the neutrons
absorbing their energy, then one may get thermal neutrons which would start the fissioning process again.
This material is called ”moderator”.
As it is shown in the Exercises, the energy output of the fission reaction is huge, of about 200 MeV.
This energy is released as kinetic energy of the outgoing particles. The machines that transform this
energy to forms which are of practical use (i. e. to thermal or electrical energy) are called ”nuclear
reactors”. In thermal power plants it is the fuel that provides heat. In the same fashion, in nuclear
reactors the fuel is the fissioning material. The nuclear fuel is usually contained in rods (”fuel rods”)
which are immersed in watter pools. The particles that are emitted in the fissioning process are absorbed
by the water, thus transforming the kinetic energy into thermal energy. At the same time, this water acts
also as moderator for the neutrons, contributing to the number of thermal neutrons that induce fission.

Chain reactions

The most common uranium isotope found in nature is 238 U and, therefore, one has to purify this isotope to
get the quantity of 235 U needed to start the fissioning reaction (12.3). This is called uranium enrichmente.
Only about 3 % of 235 U is contained in the fuel rods.
The process of fissioning through a thermal neutron hitting a fissionable material, thus generating
new thermal neutrons which start the reaction again is called ”chain reaction”. The number of neutrons
available in the chain reaction is a result of the equilibrium between the production of neutrons in the
fission process and the losses due to absorption by other materials in the fuel rods or neutrons which
leave the reactor without been scattered. When the number of neutrons which are available at the start
of the reaction is the same from one generation to the next the chain reaction is self sustained and the
reactor working condition is called ”critical”. If the production of neutrons increases from one fission
event to the next, i. e. as the reaction chain proceeds, the power level also increases, and the reactor is
said to be in a ”supercritical” condition. Finally if the production of neutrons is low, and decreases from
one generation to the next, the reactor is in a ”subcritical” condition. To get a reactor to be critical,
as all energy production reactors are, one has to use proper moderators, although also the amount of
fissionable material and the configuration of the reactor is important.
Finally, and unfortunately, supercritical devices constitute the basis of nuclear weapon production.
Besides 235 U, two other fissionable heavy nuclei used in nuclear reactors are 239 Pu and 233 U. The
process followed by these isotopes in the fission chain is the same as described above.

Theory of nuclear fission

Nuclear matter is incompressible and therefore models to explain fission treats nuclei as continuous
fluids. The most successful is the liquid drop model, in which the nucleus is represented as a fluid drop,
for instance a drop of water. According to this model the nucleus in an unstable state vibrates and
changes form from spherical to a peanut-like, or like a number 8, shape. The Coulomb repulsion between
the two lobes in the 8-shape separates them, arriving to two well differentiated spheres which finally
depart from each other, thus fissioning the mother nucleus. Usually this process occurs so fast that it
is not possible to measure any half-life. But in some cases the tunneling processes required to go from
the spherical to the fissioning shape may take a long time. Therefore fission and decay can be viewed as
similar mechanisms of decay. In fact the discovery of cluster decay mentioned at the end of last chapter,
was predicted by analyzing the decay process within a fission theory.
http://www.nature.com/nature/journal/v409/n6822/pdf/409785a0.pdf

BARRIER code: Calculation of fission barriers

http://www.sciencedirect.com/science/article/pii/S001046559900199X
http://cpc.cs.qub.ac.uk/
12 Nuclear energy and Nucleosynthesis 145

12.2 Stability of superheavy elements

It has long been a fundamental question as to what are the maximum charge and mass that a nucleus may
attain. According to classical physics, elements with Z ≥ 104 should not exist due to the large Coulomb
repulsive force. In the region of transuranium elements, the nuclei soon become unstable with increasing
atomic number. They decay very fast through α emissions and spontaneous fissions. The occurrence of
superheavy elements with Z ≥ 104 is entirely due to quantal shell effects. Theoretically, predictions have
been made that there should be an island of stability of superheavy nuclei around N = 162 or 184 and
Z = 108, 114, 120, or 126.
Recently, signicant progress has been made experimentally in the synthesis of the heaviest elements.
http://physics.aps.org/articles/v3/31

12.3 Fusion

Another form of nuclear energy is provided by fusion processes which are just the opposite to fission
reactions. In fusion two light nuclei fuse together to form a heavier nucleus, releasing also huge amount
of energy. From the energy production point of view the most promising fusion reaction is
3
H +2 H →4 He + n (12.4)

where tritium and deuteron, two isotopes of Hydrogen, fuse to form 4 He. These isotopes are found
abundantly in Nature, since water (H2 O) consists mostly of Hydrogen and Helium is a noble gas which
do not produce any contamination. Therefore if the fusion reaction would be technologically feasible, one
part of the energy problem that affects society would disappear. although the large number of neutrons
emitted in fusion reactions may still give rise to undesirable effects. But such a technology has not been
developed yet, although huge and expensive efforts are been pursued in various places around the World
(http://www.iter.org/).
Since the nuclear force is strongly attractive one may think that it should be relatively easy to fuse
two nuclei. However, to achieve fusion one has to penetrate the Coulomb barrier induced by the protons
in the nuclei, which requires very large energies and very high temperatures. This is the main difficult in
the developing of fusion reactors.
Yet, fusion reactions are very common, since it is through these reactions that stars are powered and
the elements that exist in Nature are created, as we will now study.

Heavy-ion fusion
When the incident energy is not so large and the system is not so light, the reaction process is pre-
dominantly governed by quantum tunneling over the Coulomb barrier created by the strong cancellation
between the repulsive Coulomb force and the attractive nuclear interaction. Extensive experimental as
well as theoretical studies have revealed that fusion reactions at energies near and below the Coulomb
barrier are strongly influenced by couplings of the relative motion of the colliding nuclei to several nuclear
intrinsic motions. Heavy-ion sub-barrier fusion reactions thus provide a good opportunity to address the
general problem on quantum tunneling in the presence of couplings, which has been a popular subject in
the past decade in many branches of physics and chemistry.
Theoretically the standard way to address the effects of the coupling between the relative motion and
the intrinsic degrees of freedom on fusion is to numerically solve the coupled-channel equations, including
all the relevant channels.

http://omnis.if.ufrj.br/~carlos/artigos/liqdrop1.pdf
http://iopscience.iop.org/0029-5515/47/8/001/pdf/0029-5515_47_8_001.pdf
Nucl. Fusion 47 (2007) 721727

A program for coupled-channel calculations with all order couplings

for heavy-ion fusion reactions
http://www.sciencedirect.com/science/article/pii/S001046559900243X
12 Nuclear energy and Nucleosynthesis 146

12.4 Nucleosynthesis in stars

Stars are formed through the presence of free protons in space which clump together under the influence
of the gravitational field. Free neutrons do not exist in space since, as we discussed in the previous
Chapter, they decay into protons.
In the center of the stars thus formed the protons are concentrated in a high temperature and high
density environment. In this environment the protons interact with each other to produce heavier isotopes
in a process that goes on until the protons are depleted.
The reactions that induce the creation of isotopes heavier than protons follow a rather complicated
pattern. First occurs a series of proton-proton reactions (called the pp chain), and then protons collide
with the heavier particles that are created through the pp chain, particularly Carbon, Nitrogen and
Oxygen (called the CNO cycle).

Hydrogen burning: the pp chain and CNO cycle

Our Sun belongs to the category of stars where the pp reaction chain takes place. The energy released
in these reactions is the solar energy which in our planet is responsible for everything which is alive.
As we have already mentioned, there are no free neutrons available after the formation of stars. The
question is then how the pp reactions could produce heavier elements without neutrons available to yield
deuterons. It was Bethe who realized in 1938 that nucleosynthesis in stars begins with the beta-decay
reaction
p + p −→ d + e+ + ν (12.5)
The energy released in this reaction is Q=1.44 MeV. To obtain this energy one can not use Eq. (12.2),
since now the number of neutrons and protons in the entrance channel (two protons and no neutrons)
is not the same as in the exit channel (deuteron, i. e. one neutron and one proton). One has to use
the atomic binding energies instead of the nuclear ones. As function of the mass excess the Q-value
corresponding to the reaction
a + A −→ b + B (12.6)
acquires the form.

Q = (∆Ma + ∆MA − ∆Mb − ∆MB )c2 (12.7)

The mass excesses and other related quantities are tabulated in a number of papers. In Internet these
values can be found in the same site as the nuclear binding energies.
The 1.44 MeV released by each pair of protons in the process (12.5) induces in the center of the star
huge temperatures. If one considers the star as an energy emitting black body, which is a quiet reasonable
approximation, then one would expect the energy and temperature to be related by E = kT , where
k = 8.62 × 10−11 M eV /T is the Boltzmann constant and T is degree-Kelvin. Therefore the temperature
would be T = 1.7×1010 K. But one has also to consider the reaction probability in the stellar environment
and the corresponding thermal distribution. This calculation is beyond the scope of the present course.
We will only give the final result which corresponds to the state of equilibrium at the center of the star.
This occurs at a temperature of about T = 1.6 × 107 K.
To estimate the production of deuterons in the process (12.5) one has to evaluate the corresponding
cross section, which turns out to be very small because it is determined by the weak forces in that process.
One can say that it is because the weak interaction determines the first step of the stellar nucleosynthesis
that the sun is still shining today. If instead of the weak force it was the strong force that started the
solar nucleosynthesis, the corresponding cross section would be many orders of magnitude larger and the
burning of protons would occur very fast. As a result the Sun would have died just in the beginning of
the Earth creation. On the contrary, at the present rate of hydrogen consumption, the sun will leave
about 5 billion years more, when it will become a red giant. We will come back to this subject.
Another reaction that could produce a deuteron, instead of (12.5), is

p + e− + p −→ d + ν. (12.8)

It was found that the cross section corresponding to this reaction, which is called PEP, is even smaller
than the reaction (12.5) by 4 orders of magnitude and, therefore, does not play a significant role in the
production of deuterons.
12 Nuclear energy and Nucleosynthesis 147

Deuterons in stars are then produced through the reaction (12.5). With deuterons thus formed the
next question is how these deuterons react with other particles to continue the burning process leading
to heavier elements. Of all the possible reactions having the deuteron as a target the most likely to occur
is d(p, γ)3 He. This is because in the environment of the star at this stage of nucleosynthesis the number
of protons is overwhelmingly larger than the number of any other particle. This reaction has the lowest
Coulomb barrier of all fusion reactions in the pp chain (except the weak p + p → d)
Rather involved calculations showed that the burning of 3 He is done through a chain of reactions as
follows,
• Chain I (86% of the burning)
p(p, e+ ν)d → d(p, γ)3 He →3 He(3 He, 2p)4 He
• Chain II (14% of the burning)
p(p, e+ ν)d → d(p, γ)3 He →3 He(α, γ)7 Be →7 Be(e− , ν̄)7 Li →7 Li(p, α)4 He
• Chain III (0.02% of the burning)
p(p, e+ ν)d → d(p, γ)3 He →3 He(α, γ)7 Be →7 Be(p, γ)8 B →8 B(e+ , γ)8 Be∗
→8 Be∗ (α)α
One sees that at the end of these reactions only alpha particles, i. e. Helium, is left.

CNO cycle
Besides the pp chain there is another path through which Hydrogen is burned. One of the most abundant
of the heavy elements present in the star is 12 C. Protons react with 12 C inducing a cycle as follows
12
C(p, γ)13 N →13 N (e+ , ν)13 C →13 C(p, γ)14 N →14 N (p, γ)15 O →15 O(e+ , ν)15 N →15 N (p, α)12 C
The net result of this cycle, called the CN cycle, is that four protons are converted into Helium
(Q=26.73 MeV). These four protons, which are consumed at the end of the cycle, are the ones involved
in the reactions 12 C(p, γ)13 N , 13 C(p, γ)14 N , 14 N (p, γ)15 O and 15 N (p, α)12 C. The Helium (α particle) is
created at the end of the cycle. Besides there are two positrons and two neutrinos created in this cycle.
The extraordinary feature of the CN cycle is that 12 C is only used as a catalyst. That is, it is not
consumed at all, it starts the cycle and at the end it appears again to start a new cycle. Therefore,
although its abundance is very small (about 3 × 10−6 ) it contributes significantly to the production of
alpha particles and also to the burning of protons.
Attach to the CN cycle appears another one starting in 16 O as follows,
16
O(p, γ)17 F →17 F (e+ , ν)17 O →17 O(p, α)14 N →14 N (p, γ)15 O →15 O(e+ , γ)15 N →15 N (p, γ)16 O
Once again the cycle is completed without any consumption of 16 O, which is restored at the end of
the cycle at it happened above with 12 C. As in that case, four protons are burned to produce an alpha
particle, two neutrinos and two positrons. Noticed that 16 O is produced through 15 N, which was created
in the CN cycle. But since no element is burned here the two cycles go in parallel without interruption.
The combination of these two cycles is called the CNO cycle.
The hydrogen burning of the pp chain and the CNO cycle continues until the hydrogen fuel is nearly
consumed. As a result the outgoing pressure of the radiation diminishes and the ashes, i. e. helium, are
pressed towards the center of the star. The remaining hydrogen, being much lighter than helium, forms
a layer above the helium ground. Under the gravitational force, the helium core contracts itself and,
at the same time, becomes hotter. This induces an increasing rate of pp reactions in the outer layer of
hydrogen (instead of in the center of the star, as have happened so far) which vastly raises the luminosity
of the star by a factor of 1,000 to 10,000 times. The very hot hydrogen gas on the surface of the star
expands also by very large factors and, at the same time, the outer regions of the layer becomes cooler.
This causes the spectrum of the light emitted from the star to shift towards the red, thus given rise to
the stars called red giants.
The first step after the burning of hydrogen is the burning of helium. This we will study in the next
Section.

Helium burning

When a star reaches the stage of a red giant its core consists mainly of helium. But it is very difficult to
proceed farther from helium in the path of heavier isotopes because the N=5 nuclei are not stable. The
most likely channel is
α + α → 8 Be. (12.9)
The ground state of 8 Be is unbound with respect to α decay with a Qα value of 91.84keV and a decay
width of 5.57eV. But from here no reasonable reaction was found to proceed forward. Yet, after helium
12 Nuclear energy and Nucleosynthesis 148

it is oxygen, carbon and nitrogen the most abundant elements in the Universe. The only way out of
this difficult was to assume that there was a resonance at about 7.7 MeV in 12 C such that the reaction
proceeds through the resonance by means of the reaction

α+ 8
Be → 12
C∗ (12.10)

where 12 C ∗ is the resonant state. This process is referred to as triple-α reaction. This mechanism was
proposed by Hoyle in 1953. Hoyle used the fact that 12 C is abundant in the universe as evidence for
the existence of the 12 C resonance. Since no experimental evidence of such resonance existed, Hoyle’s
idea induced some experimental efforts in that direction. One of the greatest success of nucleosynthesis
occurred in 1957 when it was experimentally found that a resonance with the exact properties needed
for the production of carbon indeed existed. The existence of such resonant states, greatly increases the
nuclear reaction rate.
Helium burning continues via the 12 C(α, γ)16 O reaction. There is no resonance near or above the
α-particle threshold in 16 O and thus this process must proceed via broadresonance tails and direct
mechanisms.
Summarizing what we have described so far, i. e. the creation of isotopes up to oxygen, we have
started with the burning of protons in the production of helium through the pp chain. This process went
on until hydrogen was near exhaustion and the pp chain was terminated. Due to its higher weight, the
helium left as ashes built a core upon which rested a layer of hydrogen. The helium core then contracted
under the influence of gravitation and high temperatures were reached. This ignited the hydrogen in the
layer, which expanded forming an enormous halo of a warm gas that irradiated light in the red region of
the spectrum. A red giant star was thus formed. The core of a red giant star consists of helium, which
under additional gravitational contraction reacts to form 8 Be and through it 12 C. The path leading to
this nucleus is very involved. It was predicted before it was corroborated experimentally in one of the
most resonant success of rather simple nuclear physics arguments. The carbon thus created, still immerse
in a see of alpha particles, reacts to create oxygen and small quantities of other nearby isotopes. This
process of helium burning proceeds up to the point when helium itself is not enough to induce additional
reactions and the helium burning process is terminated. The core consists now mainly of carbon and
oxygen. Above this heavy core there are two layers, the lowest consists of helium and the upper one,
forming the surface of the star, is hydrogen. As in previous stages of nucleosynthesis, gravitation contracts
the core farther and carbon, oxygen and, in successive steps, heavier elements start to react. This we will
analyze below.

The burning of heavy elements

The burning of heavy elements depends strongly upon the capacity of the gravitational force. If the star
is massive enough the core will contract and the temperature and density will increase until the residues
of the helium burning, i. e. carbon and oxygen ashes, starts to ignite. Since the Coulomb barrier is lower
for carbon, this will be the first to react, resulting in the formation of neon, sodium and magnesium. The
corresponding carbon burning reactions are
12 12 20 4
6 C 6 + 6 C 6 → 10 N e10 + 2 He2
12 12 23 1
6 C 6 + 6 C 6 → 11 N a12 + 1 H 0
12 12 23
6 C 6 + 6 C 6 → 12 M g 11 + n
The path followed by this carbon burning process depends upon the mass of the star. We will come
back to this point. In what follows we will assume that the mass is high enough for nucleosynthesis to
proceed farther. This continues in the same fashion as before. Thus, the production of the heavy elements
neon, sodium and magnesium in the reactions above is terminated when the fuel, i. e. carbon, is nearly
consumed, The heavy nuclei become now the core of the star. Carbon floats above this core forming the
deepest layer in the surface of the star. Gravitation induces the core to contract until a temperature
and density is reached such that high energy photons are produced. These photons interact with neon
inducing the reactions
20 16
10 N e10 + γ → +8 O 8 + α
20 24
10 N e 10 + α → 12 M g 12 + γ
After neon is consumed the core consists of magnesium and oxygen. Again this core contracts and
the oxygen burning process starts. A series of reactions takes place the result of which is that silicon
is produced. Even more complex is the mechanism induced by silicon burning. Hundreds of reactions
take place and high energy neutrinos are abundantly produced. Since these neutrinos escape the star
12 Nuclear energy and Nucleosynthesis 149

carrying with them an appreciable amount of energy, the energy output of silicon burning is not very
large. Instead, the energy produced in the burning process is spent in producing a very rapid conversion
of the core to iron and nickel. These are the most tightly bound nuclei that exist in Nature and, as a
result, they cannot participate in any reaction leading to even tighter elements, thereby releasing energy,
as it happened so far. Without this outflow of radiation to stabilize the inner pressure of gravitation, the
core collapses. In its way to its end the star undergoes a rapid succession of reactions in what is called
the ”r-process”.

The end of the star

The evolution of the star depends upon its capacity to burn different isotopes in its interior. The energy
that induces this burning is provided by the gravitational force, which compress the star and heats its
core. Since the gravitational force is determined by the mass of the star, it is this quantity that governs
the evolution of the star. Thus, a star with a mass of less than about 0.5 M , where M is the solar
mass, will be able to burn protons, but the fuel will not be enough to start the burning of helium. Due
to the low gravitation a layer of hydrogen floats on the surface of the star, where it becomes cooler and
its emitted light turns to the red part of the spectrum. These stars are called red dwarfs. They can live
very long time, since the burning of hydrogen proceeds very slowly. But eventually the hydrogen fuel
will be consumed and the star will contract under the effect of gravitation to become an small and very
compact object. The big pressure inside this stage of the star makes that the electrons present in the star
are packed together up to the limit of what the Pauli principle allows. Farther pressure will be resisted
and the collapse of the star will be prevented. This process is called ”electron degeneracy pressure”.
After shrinking to the limit allowed by electron degeneracy the outer hydrogen layer in the red dwarf
disappears and the light emitted by the small and compact object which is left turns to white. The star
becomes a white dwarf. The white dwarf will also live a very long time, emitting the thermal radiation
left after hydrogen burning. When this source of energy is also depleted, the star will become a black
dwarf. However, this process will take much more time than the age of the Universe, and therefore no
black dwarfs, or effects induced by their presence, have been observed so far.
But the white dwarf may finish in a much more dramatic fashion if another star lying nearby provides
mass. Usually this companion star is surrounded by a disk of gaseous matter which is absorbed by the
white dwarf. When the mass of the white dwarf reaches the value of 1.4 M (this is called the Chan-
drasekhar limit) the electron degeneracy pressure is overcome and the star collapses under gravitation.
An enormous explosion occurs and the mass of the star is irradiated. The radiation energy thus emitted
may be at least as high as the radiation emitted by the whole galaxy where the star belongs (depending
on the size of the galaxy). This brilliant object is called supernova. Actually, and as we will see below,
there other ways of producing supernovas. The one induced by white dwarfs by accretion of matter from
a companion star is called supernova Ia. These supernovas are excellent candles in astronomy since one
knows exactly the energy emitted by the supernova and, equally important, the spectrum of the light
emitted has a very characteristic shape and composition which makes it possible to recognize it.
Stars with masses in the range (0.5-6)M become red giants following the process that we have
described above. Given the mass of our Sun, it will finish as a red giant.
Stars with masses in the range 6 to 20 M follow the same pattern of evolution as red giants but to
a larger extend and, therefore, they become red supergiants. The interplay among all the quantities that
influence the evolution of these stars is very complicated and it may take different paths towards the end.
They may evolve as red giants do, finishing as white dwarfs. But they may also keep a heavy core that
induces a gravitational force large enough to overcome the electron degeneracy pressure. Without this
pressure the star contracts without opposition and eventually, in a period that last about one day only
(in this period occurs the r-process!) finishes as a supernova called type II.
Stars with masses up to 120M generates so much heat in their interior that the resulting radiation
pressure may blow up the whole stellar envelope. After this the star evolves into a red supergiant and
finally either a supernova or even a black hole. A black hole is an object from where, due to gravitation,
light cannot escape. It does not need to be an extremely dense object, as can be seen by calculating
its density (General Relativity and Classical Mechanics provide the same result for this number and,
therefore, its calculation is left as an Exercise). It is very difficult to predict which star would finish as a
black hole, since the mechanism of supernova collapse itself is not well understood.
Stars cannot be more massive than 120M because this is the limit where the external pressure of the
radiation equilibrates the internal pressure of gravity (this is called ”Eddington limit”). Over this limit
12 Nuclear energy and Nucleosynthesis 150

the heat generated in the center of the star produces an external pressure which overruns gravitation.
The star thus disappears before even it is formed.
When stars collapse a process of electron capture may take place. These electrons convert the protons
into neutrons, thus forming a neutral object. The Coulomb forces that keeps nucleons apart disappears
and the star becomes a very small and dense neutral object called neutron star. The neutron stars do
not collapse to become black holes because the neutrons feel the Pauli effect in a similar way as electrons
do in the electron degeneracy pressure. But they become extremely dense and small objects, with radii
of a few kilometers. As they shrink, these objects rotates more and more rapidly due to conservation of
angular momentum, reaching an angular velocity of several hundred revolutions per second. Under such
extremely rapid rotations the intrinsic magnetic moment of the star induces a very strong magnetic field
that is detected in Earth as a pulse of radiation, which gave to these stars the name of ”pulsars”. When
pulsars were first observed in 1967 it was thought that they were radio signals emitted by an advanced
extraterrestrial civilization. The researchers that found this object called it LGM-1, which means Little
Green Man number 1.
The dynamics of neutron stars are determined by the physical properties of neutrons in dense nuclear
matter. This is a subject which is at the front of nuclear physics studies at present.

http://www.int.washington.edu/PHYS554/2011/2011.html

AlterBBN: A program for calculating the BBN abundances of the elements

in alternative cosmologies
http://www.sciencedirect.com/science/article/pii/S0010465512001233

The r-process of stellar nucleosynthesis:

Astrophysics and nuclear physics
achievements and mysteries
http://arxiv.org/pdf/0705.4512.pdf

http://arxiv.org/pdf/1202.6577.pdf

Big Bang Nucleosynthesis

http://arxiv.org/abs/astro-ph/0406663
 Bibliography

[1] Rep. Prog. Phys. 74 (2011) 096901. Nuclear astrophysics: the unnished quest for the origin of the
elements, Jordi Jose, and Christian Iliadis.
Chapter 13
Constants and units

Natural units are physical units of measurement based only on universal constants. For example the
elementary charge e is a natural unit of electric charge, or the speed of light c is a natural unit of speed.
In nuclear physics, the most useful units are ~, c, fm, MeV.

Table 13.1: The values in SI units of some non-SI units based on the 2010 CODATA adjustment of the
values of the constants.
Relative std.
Quantity Symbol Numerical value Unit uncert. ur
electron volt: (e/C) J eV 1.602 176 565(35) × 10−19 J 2.2 × 10−8
atomic mass unit: 121
m(12 C) u 1.660 538 921(73) × 10−27 kg 4.4 × 10−8

Natural units (n.u.)

n.u. of velocity c, c0 299 792 458 m s−1 exact
n.u. of action: h/2π ~ 1.054 571 726(47) × 10−34 Js 4.4 × 10−8
6.582 119 28(15) × 10−16 eV s 2.2 × 10−8
~c 197.326 9718(44) MeV fm 2.2 × 10−8
n.u. of mass me 9.109 382 91(40) × 10−31 kg 4.4 × 10−8
n.u. of energy me c2 8.187 105 06(36) × 10−14 J 4.4 × 10−8
0.510 998 928(11) MeV 2.2 × 10−8
n.u. of momentum me c 2.730 924 29(12) × 10−22 kg m s−1 4.4 × 10−8
0.510 998 928(11) MeV/c 2.2 × 10−8
n.u. of length: ~/me c 386.159 268 00(25) × 10−15 m 6.5 × 10−10
n.u. of time ~/me c2
1.288 088 668 33(83) × 10−21 s 6.5 × 10−10

Atomic units (a.u.)

a.u. of charge e 1.602 176 565(35) × 10−19 C 2.2 × 10−8
a.u. of mass me 9.109 382 91(40) × 10−31 kg 4.4 × 10−8
a.u. of action: h/2π ~ 1.054 571 726(47) × 10−34 Js 4.4 × 10−8
a.u. of Bohr radius
α/4πR∞ a0 0.529 177 210 92(17) × 10−10 m 3.2 × 10−10
a.u. of Hartree energy
e2/4π0 a0 = α2 me c2 Eh 4.359 744 34(19) × 10−18 J 4.4 × 10−8
a.u. of time ~/Eh 2.418 884 326 502(12) × 10−17 s 5.0 × 10−12
a.u. of force Eh /a0 8.238 722 78(36) × 10−8 N 4.4 × 10−8
a.u. of velocity: αc a0 Eh /~ 2.187 691 263 79(71) × 106 m s−1 3.2 × 10−10
a.u. of momentum ~/a0 1.992 851 740(88) × 10−24 kg m s−1 4.4 × 10−8
a.u. of current eEh /~ 6.623 617 95(15) × 10−3 A 2.2 × 10−8
a.u. of charge density e/a30 1.081 202 338(24) × 1012 C m−3 2.2 × 10−8
a.u. of electric potential Eh /e 27.211 385 05(60) V 2.2 × 10−8
a.u. of electric field Eh /ea0 5.142 206 52(11) × 1011 V m−1 2.2 × 10−8
a.u. of el dipole mom ea0 8.478 353 26(19) × 10−30 Cm 2.2 × 10−8
a.u. of el quadrupole mom ea20 4.486 551 331(99) × 10−40 C m2 2.2 × 10−8

The Bohr radius approximately equals to the most probable distance between the proton and electron
in a hydrogen atom in its ground state.
4πε0 ~2 ~ α
a0 = 2
= = , (13.1)
me e me cα 4πR∞
where ε0 is the permittivity of free space. The Rydberg constant is defined as R∞ = α2 me c/2h. A
muonic hydrogen atom consists of a negative muon and a proton. Since mµ /me ≈ 207, the Bohr radius
of the muon is about 200 times smaller than the electron Bohr radius.
 13

1 12
Table 13.2: The values of some energy equivalents derived from the relations E = mc2 = hc/λ = hν = kT ; 1 eV = (e/C) J, 1 u = mu = 12 m( C) =
10−3 kg mol−1/NA .
Relevant unit
Constants and units

J kg m−1 Hz

1J (1 J) = (1 J)/c2 = (1 J)/hc = (1 J)/h =

1J 1.112 650 056 . . . × 10−17 kg 5.034 117 01(22) × 1024 m−1 1.509 190 311(67) × 1033 Hz

1 kg (1 kg)c2 = (1 kg) = (1 kg)c/h = (1 kg)c2 /h =

8.987 551 787 . . . × 1016 J 1 kg 4.524 438 73(20) × 1041 m−1 1.356 392 608(60) × 1050 Hz

1 m−1 (1 m−1 )hc = (1 m−1 )h/c = (1 m−1 ) = (1 m−1 )c =

1.986 445 684(88) × 10−25 J 2.210 218 902(98) × 10−42 kg 1 m−1 299 792 458 Hz

1 Hz (1 Hz)h = (1 Hz)h/c2 = (1 Hz)/c = (1 Hz) =

6.626 069 57(29) × 10−34 J 7.372 496 68(33) × 10−51 kg 3.335 640 951 . . . × 10−9 m−1 1 Hz

1K (1 K)k = (1 K)k/c2 = (1 K)k/hc = (1 K)k/h =

1.380 6488(13) × 10−23 J 1.536 1790(14) × 10−40 kg 69.503 476(63) m−1 2.083 6618(19) × 1010 Hz

1 eV (1 eV) = (1 eV)/c2 = (1 eV)/hc = (1 eV)/h =

1.602 176 565(35) × 10−19 J 1.782 661 845(39) × 10−36 kg 8.065 544 29(18) × 105 m−1 2.417 989 348(53) × 1014 Hz

1u (1 u)c2 = (1 u) = (1 u)c/h = (1 u)c2 /h =

1.492 417 954(66) × 10−10 J 1.660 538 921(73) × 10−27 kg 7.513 006 6042(53) × 1014 m−1 2.252 342 7168(16) × 1023 Hz

1 Eh (1 Eh ) = (1 Eh )/c2 = (1 Eh )/hc = (1 Eh )/h =

4.359 744 34(19) × 10−18 J 4.850 869 79(21) × 10−35 kg 2.194 746 313 708(11) × 107 m−1 6.579 683 920 729(33) × 1015 Hz
153
 13

1 12
Table 13.3: The values of some energy equivalents derived from the relations E = mc2 = hc/λ = hν = kT ; 1 eV = (e/C) J, 1 u = mu = 12 m( C) =
10−3 kg mol−1/NA .
Relevant unit
Constants and units

K eV u Eh

1J (1 J)/k = (1 J) = (1 J)/c2 = (1 J) =
7.242 9716(66) × 1022 K 6.241 509 34(14) × 1018 eV 6.700 535 85(30) × 109 u 2.293 712 48(10) × 1017 Eh

1 kg (1 kg)c2 /k = (1 kg)c2 = (1 kg) = (1 kg)c2 =

6.509 6582(59) × 1039 K 5.609 588 85(12) × 1035 eV 6.022 141 29(27) × 1026 u 2.061 485 968(91) × 1034 Eh

1 m−1 (1 m−1 )hc/k = (1 m−1 )hc = (1 m−1 )h/c = (1 m−1 )hc =

1.438 7770(13) × 10−2 K 1.239 841 930(27) × 10−6 eV 1.331 025 051 20(94) × 10−15 u 4.556 335 252 755(23) × 10−8 Eh

1 Hz (1 Hz)h/k = (1 Hz)h = (1 Hz)h/c2 = (1 Hz)h =

4.799 2434(44) × 10−11 K 4.135 667 516(91) × 10−15 eV 4.439 821 6689(31) × 10−24 u 1.519 829 846 0045(76) × 10−16 Eh

1K (1 K) = (1 K)k = (1 K)k/c2 = (1 K)k =

1K 8.617 3324(78) × 10−5 eV 9.251 0868(84) × 10−14 u 3.166 8114(29) × 10−6 Eh

1 eV (1 eV)/k = (1 eV) = (1 eV)/c2 = (1 eV) =

1.160 4519(11) × 104 K 1 eV 1.073 544 150(24) × 10−9 u 3.674 932 379(81) × 10−2 Eh

1u (1 u)c2 /k = (1 u)c2 = (1 u) = (1 u)c2 =

1.080 954 08(98) × 1013 K 931.494 061(21) × 106 eV 1u 3.423 177 6845(24) × 107 Eh

1 Eh (1 Eh )/k = (1 Eh ) = (1 Eh )/c2 = (1 Eh ) =
3.157 7504(29) × 105 K 27.211 385 05(60) eV 2.921 262 3246(21) × 10−8 u 1 Eh
154
13 Constants and units 155

Table 13.4: Some exact and measured quantities.

Quantity Symbol Value

speed of light in vacuum c, c0 299 792 458 m s−1
magnetic constant µ0 4π10−7 N A−2 = 12.566 370 614 × 10−7 N A−2
electric constant 0 (µ0 c2 )−1 = 8.854 187 817... × 10−12 F m−1
Newtonian constant G 6.673 84(80) × 10−11 m3 kg−1 s−2
of gravitation G/~c 6.708 37(80) × 10−39 (GeV/c2 )−2
Planck constant h 6.626 069 57(29) × 10−34 J s
4.135 667 516(91) × 10−15 eV s
h/2π ~ 1.054 571 726(47) × 10−34 J s
6.582 119 28(15) × 10−16 eV s
~c 197.326 9718(44) MeV fm
Planck mass (~c/G)1/2 mP 2.176 51(13) × 10−8 kg
energy equivalent mP c2 1.220 932(73) × 1019 GeV
Planck temperature (~c5 /G)1/2 /k TP 1.416 833(85) × 1032 K
Planck length ~/mP c = (~G/c3 )1/2 lP 1.616 199(97) × 10−35 m
Planck time lP /c = (~G/c5 )1/2 tP 5.391 06(32) × 10−44 s
elementary charge e 1.602 176 565(35) × 10−19 C
e/h 2.417 989 348(53) × 1014 A J−1
Bohr magneton e~/2me µB 927.400 968(20) × 10−26 J T−1
5.788 381 8066(38) × 10−5 eV T−1
µB /h 13.996 245 55(31) × 109 Hz T−1
µB /hc 46.686 4498(10) m−1 T−1
µB /k 0.671 713 88(61) K T−1
nuclear magneton e~/2mp µN 5.050 783 53(11) × 10−27 J T−1
3.152 451 2605(22) × 10−8 eV T−1
µN /h 7.622 593 57(17) MHz T−1
µN /hc 2.542 623 527(56) × 10−2 m−1 T−1
µN /k 3.658 2682(33) × 10−4 K T−1
molar mass of 12 C M (12 C) 12 × 10−3 kg mol−1
molar mass constant Mu 10−3 kg mol−1
relative atomic mass of 12 C Ar (12 C) 12
electron relative atomic mass Ar (e) 0.000 548 579 9111(12)
neutron relative atomic mass 1.008 664 915 74(56)
1
H relative atomic mass 1.007 825 032 07(10)
2
H relative atomic mass 2.014 101 777 85(36)
4
He relative atomic mass 4.002 603 254 153(63)
electron mass me 9.109 382 91(40) × 10−31 kg
5.485 799 0946(22) × 10−4 u
energy equivalent me c2 8.187 105 06(36) × 10−14 J
0.510 998 928(11) MeV
electron-muon mass ratio me /mµ 4.836 331 66(12) × 10−3
electron-tau mass ratio me /mτ 2.875 92(26) × 10−4
electron-proton mass ratio me /mp 5.446 170 2178(22) × 10−4
electron-neutron mass ratio me /mn 5.438 673 4461(32) × 10−4
proton mass mp 1.672 621 777(74) × 10−27 kg
1.007 276 466 812(90) u
energy equivalent mp c2 1.503 277 484(66) × 10−10 J
938.272 046(21) MeV
proton-electron mass ratio mp /me 1836.152 672 45(75)
proton-neutron mass ratio mp /mn 0.998 623 478 26(45)
neutron mass mn 1.674 927 351(74) × 10−27 kg
1.008 664 916 00(43) u
energy equivalent mn c2 1.505 349 631(66) × 10−10 J
939.565 379(21) MeV
alpha particle mass mα 6.644 656 75(29) × 10−27 kg
13 Constants and units 156

Table 13.4: (Continued).

Quantity Symbol Value

4.001 506 179 125(62) u
energy equivalent mα c2 5.971 919 67(26) × 10−10 J
3727.379 240(82) MeV
fine-structure constant e2/4π0 ~c α 7.297 352 5698(24) × 10−3
inverse fine-structure constant α−1 137.035 999 074(44)
Bohr radius α/4πR∞ = 4π0 ~2/me e2 a0 0.529 177 210 92(17) × 10−10 m
electron magnetic moment µe −928.476 430(21) × 10−26 J T−1
to Bohr magneton ratio µe /µB −1.001 159 652 180 76(27)
to nuclear magneton ratio µe /µN −1838.281 970 90(75)
proton rms charge radius rp 0.8775(51) × 10−15 m
deuteron rms charge radius rd 2.1424(21) × 10−15 m
proton magnetic moment µp 1.410 606 743(33) × 10−26 J T−1
to Bohr magneton ratio µp /µB 1.521 032 210(12) × 10−3
to nuclear magneton ratio µp /µN 2.792 847 356(23)
proton g-factor 2µp /µN gp 5.585 694 713(46)
p-n magnetic moment ratio µp /µn −1.459 898 06(34)
neutron g-factor 2µn /µN gn −3.826 085 45(90)
triton g-factor 2µt /µN gt 5.957 924 896(76)

The mass of the tau lepton mτ , the Fermi coupling constant GF and the mass ratio of the W± and
0
Z bosons obtained from the most recent report of the Particle Data Group are

mτ c2 = 1776.82(16) MeV (13.2)

GF
= 1.166 364(5) × 10−5 GeV−2 (13.3)
(~c)3
mW
= 0.8819(12) (13.4)
mZ
http://physics.nist.gov/cuu/Constants/index.html
http://arxiv.org/abs/1203.5425
http://pdg.web.cern.ch/pdg/