The Nuclear Shell Model
What is a shell model?
A shell model is one in which the system is thought to consist of individual
particles moving in bound orbitals in response to the remainder of the
system. Each orbital has a well designated energy, angular momentum,
and parity associated with it. In an atom, the electrons are bound
to the highly charged nucleus, which contains most of the mass of the atom,
by the electric force. Since most of the mass is contained in the
nucleus, the electrons can move in orbits relatively free of any obstacles
and hence would suffer very few collisions in their eternal orbiting about
the nucleus. In such a model, treated quantum-mechanically, no two
electrons, two protons, or two neutrons can occupy the same quantum state,
ie have identical sets of quantum numbers. This principle attributed
to Pauli results in a finite number of such particles occupying a given
energy level, and thus, leads to the concept of closed (or filled) levels
(or shells). When a shell is filled, any additional particles of
that type must be put in a different level (shell).
Conceptual problem with a nuclear shell model
For nuclei we not a saturation of nuclear forces resulting in an approximate
constant binding energy for each constituent nucleon, independent of the
details of nuclear structure. This is attributed to the fact that
the size of the nucleus is basically proportional to the number of nucleons
and hence the nucleus seems to be a rather compact object with nucleons
basically touching each other. Thus, one might
question the validity of treating a nucleon as if it were moving in an
orbital through this dense nuclear medium, keeping itself in a state of
constant energy implying that collisions with the other nucleons were very
infrequent. The Pauli Exclusion Principle
perhaps provides a little philosophical support for proceeding with development
of such a model. Because the different nucleons must enter different
energy and angular momentum states this perhaps forces the orbits to be
enough different that all the nucleons can move in a harmonious fashion,
suffering collisions frequently, but still remaining in discrete orbitals
enough of the time to make the model have some validity.
Experimental evidence to support the shell model
approach
Several kinds of experimental data support the concept
that nuclei demonstrate a periodic structure, just like the elements, indicating
that one major energy shell with particular energies, angular momenta,
and parities can be filled, thus forcing an abrupt change in observed quantities.
The data indicate that closed shells occur at neutron or proton numbers
of 2,8,20,28,50,82,
and 126 (called
Magic Numbers).
The types of data showing discontinuities at these magic numbers indicating
a shell closure are discussed in Krane and include some of the following:
-
Proton and Neutron Separation Energies
-
Nuclear Spin and Parity Systematics
-
Nuclear Binding Energy (increased stability of closed
shell nuclei)
-
Small variations of nuclear radii from the R = roA1/3
law.
-
Alpha Decay Energies and Half-Life Systematics
-
Neutron Absorption Cross Sections
-
Magnetic Dipole and Electric Quadrupole Moments
The accumulation of data suggesting shell structure
encourage us to pursue such a model, even in the face of the conceptual
problem with the probable frequent collisions of nucleons in the nucleus.
The Nuclear Shell Model Potential
The nuclear shell model potential should represent
the average interaction of all the other nucleons exerted on any one of
the nucleons in the nucleus. One can make quantum-mechanical calculations
using square-well potentials and spherical harmonic oscillator potentials,
with parameters adjusted to try to reproduce the binding energy of the
nucleon and the proper nuclear radius. Since the potential represents
the average interaction with the other nucleons, and since the nucleon-nucleon
interaction is short range, we might do better if we would use a potential
that mimics the mass distribution of the nucleus. Hence, modern shell
model calculations frequently use the Woods-Saxon form of the shell model
potential.
This potential is relatively flat inside the nucleus and falls gradually
to zero at the nuclear surface. For a nucleus with 100 nucleons it
has the following general shape:
Neither the infinite square well potential, the spherical harmonic oscillator
potential, nor the Woods-Saxon produces energy eigenvalues with closed
levels (shells) for all the magic numbers. In particular the levels
predicted by the model using the Woods-Saxon treatment is shown on the
left side of Fig. 5.6 on p. 123 of Krane. The lowest level is the
1s state which has only 1 ml substate (ml = 0) and
two ms values to give a total degeneracy of 2. Thus, two
protons or two neutrons can fit into this lowest level. A third neutron
will have to go into the second level which is somewhat higher in energy,
thus agreeing with the experimental fact that there is a major shell filling
at 2. The second level is the 1p state which has 3 possible ml
values, and hence a total degeneracy of 6, ie 6 protons and 6 neutrons
can occupy the 1p level. When it is filled there will be a total
of 8 protons and 8 neutrons in the nucleus. Placement of another
nucleon would require that it be placed in the next highest level, or the
1d level. Thus, the model predicts a magic number at 8 in agreement
with observation. The 1d and 2s levels are predicted to be near each
other, being almost degenerate. These two levels can accomodate 12
neutrons and 12 protons, which bring us to the next major shell closure
at 20. Above Z or N = 20, the prediction of closed shells by this
model breaks down; there are no level fillings at 28, 50, 82, or 126.
Spin Orbit Forces to the Rescue
In 1949 Maria Goeppert Mayer and independently, Haxel, Suess, and Jensen
had a brilliant idea to split some of the energy levels by applying a small
term to the potential which depends on the relative orientation of the
spin of the nucleon with regard to the direction of the orbital angular
momentum, ie a s.l or spin-orbit term. The total potential
is then V(r) = VWS(r) + VSO(r)(l.s).
For a given l there are two values of j, ie j = l + 1/2 and
j = l - 1/2, and (l.s) is positive for the first case and
negative for the second case. Thus, the energy level is lowered for
the nucleons in l + 1/2 states and raised for the l - 1/2
states. The splitting of a level is proportional to 2l+1 and
is, thus, larger for the larger l states. On the right side
of Fig. 5.6 on page 123 the resulting level diagram is shown; now that
ml and ms are not good quantum numbers, one
must count mj values to get the degeneracy of each level.
It is observed that the other major shells at 28, 50, 82, and 126 are not
reproduced and the theory predicts the next magic number will be 184.
The inclusion of the spin-orbit force was effective in reproducing all
the major closed shells (magic numbers).
Side Proof of Dependence of Level Splitting on l
-
The total angular momentum of a nucleon is the sum of the orbital and spin
angular momenta, thus j = l + s.
-
Take the dot product: j.j = (l+s).(l+s)
= l.l + s.s + 2 l.s, or j2
= l2 + s2 +2(l.s).
-
Solve for the dot product: l.s = 1/2(j2
- l2 - s2 )
-
Take the expectation value: <l.s> = hbar2/2[j(j+1)
- l(l+1) -s(s+1)]
-
Now calculate the difference in <l.s> for the j =
l + 1/2 and j = l - 1/2 levels. <l.s>j
= l + 1/2 - <l.s>j = l - 1/2
= hbar/2[(l+1/2)(l+3/2)-l(l+1)-s(s+1)-(l-1/2)(l+1/2)+l(l+1)+s(s+1)]
= hbar/2(l+1/2)(l+3/2-l+1/2) = (hbar/2)(2l+1).
Extreme Single Particle Shell Model
In the extreme single particle model, one assumes that the ground state
of an odd-A nucleus is determined by the quantum numbers of the odd nucleon
after the A-1 other nucleons have coupled to spin 0. Thus, 40Ca
with 20 protons and 20 neutrons would have all its nucleons paired and
the Ip should be 0+.
Indeed, the table in the back of Krane reports the experimental Ip
is 0+. By looking at the level predictions in Fig. 5.6
one would expect the next neutron to go into the f7/2 level
and hence, she would expect the Ip
of 41Ca to be 7/2-. Likewise, since the last
level filling before N = 20 was the d3/2 level, we would expect
the ground state of 39Ca to have an of 3/2+.
A glimpse in the appendix shows that these expectations are satisfied
by experimental values. Another point of interest is the presence
of two excited states in 41Ca at an energy of about 2 MeV which
have Ip of 3/2+
and 3/2-. These first of these states is produced by raising
one neutron from the d3/2 state to form a pair in the f7/2
level but leaving an unpaired neutron in the d3/2 state.
The second is produced by raising the odd neutron from the f7/2
level to the p3/2 level. These data indicate that
the energy gap between the d3/2 state and the f7/2
level is about 2 MeV, and the gap between the f7/2 and p3/2
level is similar.(To be precise we would need to consider the difference
in the pairing energy for two neutrons in the f and d states.) Such
observations seem to justify the shell model approach for nuclei near the
closed shells.
Magnetic Moments in the Extreme Single Particle Model
The magnetic moment of a nucleon in an orbit has orbital and intrinsic
parts. The total moment is predicted to be different for j
= l + 1/2 and j = l - 1/2 levels because of the
difference in coupling of the orbital and spin parts. The single-particle
predictions are shown as solid lines in Fig. 5.9, page 127 and are called
Schmidt limits. It is noted that most of the measured magnetic moments
are smaller than the Schmidt limits. Recalling that the intrinsic
component of the magnetic moment depends on the charge and current distributions
within the nucleon. We conclude that when other nucleons are present,
the charge distribution of the interacting nucleons is affected, thus changing
the intrinsic moment when the nucleon is bound to other nucleons to form
a nucleus. If one reduces the gs for bound nucleons to
be 0.6 of the gs for free nucleons, then the Schmidt predictions
are the dashed lines in Fig. 5.9, which do indeed kind of trace the experimental
data. This result suggests that the polarizing effects is such as to make
gs(bound) = 0.6 gs(free).
Electric Quadrupole Moments in the Single Particle Model
Electric Quadrupole Moments are not predicted very well by the spherical
shell model. For nuclei near closed shells, the calculations tend
to give the correct sign, but the magnitude is usually missed by factors
of about 2 or 3. Away from closed shells where several valence nucleons
may occupy a particular orbital, the quadrupole moments become very large
demonstrating large collective effects in nuclear motion.
What happens as you move away from a closed shell?
Although the extreme single particle model works well near the closed shell,
we quickly encounter problems as we move to several nucleons away from
the major closed shell. The assumption that all but the one odd particle
pair to 0 angular momentum is too simple when we consider possible excited
states. For example, a good demonstration of this complication is
shown in Fig. 5.12 on page 132. 41Ca and 41Sc
each has one nucleon beyond a double closed shell nucleus 40Ca.
The ground states are 7/2- indicating that the 1 valence nucleon
is in the f7/2 orbital, which agrees with the model predictions.
Note that both have excited p3/2 and d3/2 levels
around 2 MeV, as you might expect if the odd f7/2 nucleon were
moved to the adjacent shell model orbitals. So the single particle
model looks pretty good.
However, notice what happens when two additional neutrons are added
to make 43Ca and 43Sc. The ground states are
still 7/2- as we would expect, but the excited states are completely
changed. So, the result is that we must expand the model to consider
all the nucleons beyond the major closed shell as valence nucleons and
consider all the different ways these valence nucleons can couple to make
states. Such shell model calculations become quite formidable and
require super computers for numerical predictions of shell model properties.
For heavy nuclei, the shells contain too many possible orbitals to make
detailed calculations, even with the most powerful computers. Currently,
nuclear theorists are devising many clever techniques to overcome these
computational complications, including two in our own department, Drs.
David Dean and Yang Sun.