Nuclear and Particle Physics - Lecture 19

The semi-empirical mass formula

1 Introduction
The liquid drop model reproduces the gross features of the binding energy but does not give the
dependences on the individual numbers of protons and neutrons. For this we need to add more
terms which result in the “semi-empirical mass formula”.

2 The semi-empirical mass formula

We saw the binding energy can be expressed as
BE = av A − as A2/3
where av (the volume term constant) and as (the surface term constant) are parameters to be
determined from data.
The above expression for the binding energy is a reasonable approximation but does not
reproduce a lot of the effects we have already seen. It predicts that the ground state binding
energy only depends on A but is independent of Z or N , which we have already seen is not
correct. It ignors quantum effects such as the Pauli exclusion principle (and others) completely.
To get a better agreement we need to add three more terms which explicitly depend on Z and
N.
The first term we will now add is easy to understand. We know there is an EM repulsive
force between protons due to their charge and so this will reduce the binding energy for nucleons
with several protons. As we believe the nuclear force itself is independent of nucleon type, then
the protons will on average be spread evenly throughout the nucleus, which means the charge
density is uniform. It is a standard problem in electrostatics to calculate the energy required to
assemble a sphere of uniform charge density and the result is
3 Q2
5 4π0 r
For a nucleus with Z protons, this self-energy is therefore
3 e2 Z2 Z2
∆BE = − = −a c
5 4π0 r0 A1/3 A1/3
where ac = 0.72 MeV. Note the dependence on Z 2 , not Z; the EM force is long range and so
every proton affects every other proton in the nucleus, not just its nearest neighbours. Contrast
this to the short range nuclear force where the nucleons only affect their nearest neighbours and
the energy depends on A. In fact, although this Z 2 form is often used, strictly speaking it is not
quite right. The energy given by the expression above is that needed to spread all the charge out
throughout all space to an infinitely small density. However, the binding energy is defined as the
energy need to break the nucleus into its constituent nucleons, i.e. break it into neutrons and
protons, but not to spread the individual proton charges out. Indeed, the equation says even
one proton, i.e. Z = 1, gives a correction to the binding energy, even though there is nothing to
repel it. This means the correction to the binding energy should not be quite as big. A better
form is ac Z(Z − 1)/A1/3 , which is now zero for Z = 1. Clearly, for large Z, as found in large
nuclei, these two are very similar. We now have
Z(Z − 1)
BE = av A − as A2/3 − ac
A1/3

1
This by itself breaks the independence on Z but clearly predicts the biggest binding energy for
any A will be for Z = 0 or 1. We know this isn’t right as we have stable nuclei with all values of
charge up to Z > 100 in the atomic periodic table. We need to add two more terms to account
for quantum effects.
The next term we will consider is called the “asymmetry term”. The idea here is identical
to the concept of a Fermi level in the physics of materials. The nucleons have energy levels in
the nucleus and, being spin 1/2 particles, then each level can take two of each type of nucleon.

p n

If we really tried to form a nucleus purely from neutrons, as implied by the terms we have so
far for the binding energy, they would have to be put into higher and higher energy levels and
so would be less and less strongly bound, reducing the binding energy. Clearly, putting protons
into the nucleus instead would be beneficial for the binding energy as they could go into the
deepest empty proton levels. It is clear the best situation is when the two are evenly balanced
with N = Z. The details of the exact energy levels and numbers per level will be messy and
vary with A, but a reasonable parametrisation turns out to given by ∆BE ∝ −(N − Z)2 , i.e.
the binding energy is reduced symmetrically for either N > Z or Z > N . In fact, the spacing
between states depends inversely on the size of the nucleus (as it does for a square well potential)
so that larger nuclei have less of a binding energy loss if N 6= Z, hence, the full term used is
−aa (N − Z)2 /A.
The final term which is needed is called the “pairing term”. This occurs because of the
different overlap of wavefunctions for pairs of nuclei in various states. For two identical nucleons
in the same spatial state, with opposite spins to be antisymmetric as required, then the spatial
wavefunctions are effectively identical and have maximal overlap. Because of the short range
force, this gives more of a binding energy for this particular pair. This effect occurs for all
nucleons except potentially the ones in the highest occupied energy level for each type of nucleon,
where there is either one or two nucleons of that type. Hence, the nucleus will be more strongly
bound for ones with an even number of nucleons of either type. There are three cases
1. Even-even, meaning an even number of both protons and neutrons, and hence even A.
This has both pairs strongly bound.

2. Odd-odd, meaning an odd number of both protons and neutrons, and hence also even A.
This is the least strongly bound.

3. Even-odd, meaning an even number of one type and an odd number of the other, and
hence odd A. This has one strongly bound pair and so should be half way in between the
previous two.
This is therefore simply parametrised by a form ap /A1/2 , where ap is takes its positive value for
even-even nuclei, its negative value for odd-odd nuclei and is zero for even-odd nuclei. Note, the

2
pairing term implies even-even nuclei always have the spins of the nucleons in the same spatial
state parallel, so all such nuclei would be expected to have ground states with total spin zero;
this is observed to be true.
Therefore, the total expression for the binding energy is

Z(Z − 1) (N − Z)2 1
BE = av A − as A2/3 − ac − a a + ap 1/2
A 1/3 A A
and this is the semi-empirical mass formula. The best-fit parameters take values around a v =
15.8 MeV, as = 18.3 MeV, ac = 0.71 MeV, aa = 23.2 MeV and ap = ±11.2 MeV.

3 The beta-stability curve

The semi-empirical mass formula is a function of two variables, as A = N + Z. It gives the
binding energy of the ground state of any nucleus, i.e. any values of Z and N . However, we
have not observed nuclei with most of the combinations of Z and N which might be thought
possible in the whole of the Z,N plane because the majority of them are highly unstable. The
binding energy is largest in a specific region of the Z,N plane which is called the beta-stability
valley (“valley” as the mass is at a minimum, “beta-stable” for reasons which will become clear
in a later lecture).
It is easiest to analyse the binding energy along lines of constant A and we will look at it as
a function of Z. Putting N = A − Z, then
ap ac ac 4aa
     
2/3
BE = a v A − a s A − aa A + 1/2 + + 4aa Z + − 1/3 − Z2
A A 1/3 A A
For odd A, when the pairing term ap is zero, then this is a quadratic
BE

Odd A

There is a particular value of Z for which the binding energy is maximum, but away from this
value, the binding energy falls and eventually goes negative at which point the nucleus is no
longer bound. Even within the region of positive binding energy, the non-maximum values of Z
are not necessarily stable; as we will see, beta decay allows them to change protons to neutrons
and vice versa, and so move along the curve to the maximum value. Hence, we only see the
reasonably long-lived nuclei which are at or near the maximum.
For even A, then as Z changes by one, then Z (and N , given that A is fixed and even) goes
from even to odd or vice versa. This means the pairing term changes sign. Hence, this shifts the
quadratic curve up and down by ±ap /A1/2 for the alternating even and odd Z values. However,
the same stability arguments still hold; only nuclei near the peak live long enough to be seen.

3
 BE

Even A

Even Z

Z
Odd Z

We need to know how the Z position of the peak changes with A. In principle we can approximate
Z as continuous and take derivatives of the above quadratic function to find the maximum as
a function of A. However, it is also useful to get an intuitive feel for what we would expect.
The two terms driving this are the Coulomb and asymmetry terms; only ac and aa appear in
the Z and Z 2 terms of the quadratic. If ac = 0, then clearly N = Z = A/2 gives the maximum
binding energy as the nuclei like to have equal numbers of levels filled. Conversely, if aa = 0,
then Z = 0 or 1 to minimise the Coloumb term and hence maximise the binding energy by
limiting the amount of Coulomb repulsion. Hence, with both terms, we would generally expect
that Z would be somewhat less than A/2 and N would be more than A/2. The relative size
of these two terms is not the same for all A. The asymmetry term falls off as 1/A while the
Coulomb terms falls off only as 1/A1/3 . Hence the latter becomes more important as A increases.
Conversely, for small A, particularly as ac ≈ 0.7 MeV  aa ≈ 23.2 MeV, the Coulomb term has
little effect. Therefore, we expect Z ≈ A/2 for small A but Z < A/2 for large A. The actual
stable and observed nuclei which form the beta-stability valley are shown below

4
The line of the most stable nuclei is call the beta-stability curve. Note, there can be several stable
nuclei for a given A. Nuclei lying on the beta-stability curve are often studied. For example,
the binding energy per nucleon of these as a function of A are shown below and compared with
the semi-empirical mass formula.

The agreement is generally very good. Note, the binding energy per nucleon is reasonably
constant, between 7.8 and 8.8 MeV for all nuclei with A > 30. It peaks around 56 26 Fe, which
means this is the most strongly bound nucleus. It is therefore energetically favourable both to
break up heavier nuclei to bring them closer to Fe and also to combine lighter nuclei. The former
is called nuclear fission and the latter nuclear fusion. The energy released in both cases can and
has been used both for power and nuclear weapons.
The small peaks showing disagreement are a result of quantum energy levels at particular
values of Z and N and are the subject of the next lecture.