Flavour Symmetry in Particle Physics

Akash Jain
Department of Physics, IISER Bhopal
November 15, 2013
Abstract
The family of Hadrons was a complete mess until mid 1900s, when Gell-mann intro-
duced his Eightfold-Way Diagrams and brought order to their clan. The diagrams might
appear to be just a fortunate chance, but they come naturally in the representation the-
ory of SU
3
. In fact complete theory of Hadron Structure can be developed just from the
argument of symmetry and group theory. My aim through this Term Paper is to establish
this relationship between Group Theory and Particle Physics, and to review the theory for
observed patterns in Hadrons.
1 Introduction
In 1917 Emmy Noether gave her famous theorem stating - For every Conservation Law there
exists an underlying Symmetry and vice-versa. The theorem implies that all conserved quan-
tities we encounter in physics, say momentum, energy, angular momentum or charge, can be
thought of as a result of an underlying mathematical symmetry in the action of the physical
system under consideration. Conversely, if we are able to identify a conserved quantity in our
system, there must exist an underlying symmetry. Symmetries thus play a crucial role in our
understanding of physics. Much information about the system can be extracted just from the
argument of symmetry, even when a complete dynamical theory is not available.
Symmetry, in precise language, is a mathematical operation which leaves the system invariant.
In Physics one generally encounters two kinds of symmetries - Continuous and Discrete. As
name suggests Continuous Symmetries are operations which can be reached from one state to
another by continuous set of transformations, like Lorentz Transformations; on the other hand
Discrete Symmetries can not be reached by continuous transformations, like Parity.
Another distinction within Symmetries that is quite common in Physics is Global and Local
Symmetry. When a mathematical operation (which leaves the system invariant) acts on all
space-time uniformly, it is called a Global Symmetry. Whereas, if the operation is space-time
dependent, it is called a Local Symmetry. For example spatial rotations and translations are
Global Symmetries, whereas twists and expansions of space-time are Local Symmetries. In this
write-up I will restrict myself to Global Continuous Symmetries, and especially their applications
in Particle Physics.
1
2 Group Theory
Group Theory is a branch of mathematics used to study symmetries. I shall review here quickly
some of its aspects which we will be using in subsequent sections. An extensive literature on the
topic can be found in any standard text on Group Theory or Mathematical Physics[7][8].
2.1 Groups
An abstract Group G is a set of elements (nite or innite) with an operation () dened on it,
which obeys:
1. u, v G, u v G
2. u, v, w G, u (v w) = (u v) w
3. I G s.t. u I = I u = u u G
4. u G, u
1
G s.t. u u
1
= I
Readers can check that the set of all possible spatial rotations under operation succession (u w
is equivalent to rotation u succeeded by w) forms a group.
Continuous and Discrete Groups: A group whose all elements can be reached by a con-
tinuous change of one or more parameters (say, angle) is called a Continuous Group or a Lie
Group. Otherwise if the elements of the group can be labelled by integral indices, it is called a
Discrete Group. Examples of Lie Groups includes set of all real numbers over usual addition
and set of orientation preserving rotations, whereas set of all even integers over addition
and Parity-Identity pair of rotations (A group with only two elements - Identity and Parity
operator) are examples of discrete groups.
2.2 Group Representations
Every group can be represented in form of a group of invertible matrices over multiplication.
This mapping is called a representation of the group. A representation D of group G is given
by:
D : G D(G) GL
n
(F)
Where n is called the dimension of representation and F is the eld over which matrices are
dened (like R or C). GL
n
(F) is the set of all invertible matrices.
Faithful Representations: Representations need not be 1-1; most trivial representation can
thus be a map sending all the group elements to I matrix. A 1-1 representation, commonly
called Faithful Representation, contains all the information about the original group. O
3
(A GL
3
(R) | A
T
A = I) is a faithful representation for all spatial rotations, whereas SO
3
(A SO
3
| det(A) = 1) is a faithful representation for all orientation preserving rotations.
Fundamental Representations: A matrix group can itself be represented in form of other
matrix groups. For example SO
3
is a non-faithful representation of SU
2
(A GL
2
(R) | A

A =
I, det(A) = 1). A representation formed by mapping a matrix group to itself (or another matrix
group of same dimension) is called a Fundamental Representation of the group.
Actually corresponding to every element in SO
3
there are 2 elements in SU
2
, thus SU
2
is called
(by abuse of language) a double valued representation of SO
3
.
2
Irreducible Representations: Given two representations D
1
(G) and D
2
(G) of a group G,
you can always form a new representation by arranging them in block diagonal form:
D(G) =

D
1
(G) 0
0 D
2
(G)

This is fundamentally not a new representation, so we restrict ourselves to Irreducible Repre-

sentations dened by - Representations which cannot be reduced to a block diagonal form by a
similarity transformation
1
.
2.3 Lie Groups and Lie Algebras
As I have mentioned earlier, Groups whose elements can be reached by smooth variation of
some continuous parameter(s) are called Lie Groups. Thus to study their properties, it suces
to look at innitesimal transformations near the identity, as we can always integrate many such
transformations to get a nite element of the group.
Consider G, representation of a lie group H, whose elements are given by G(
1
,
2
, ) G
dened smoothly on some continuous parameters
1
,
2
, such that u(
i
= 0) = I. If we
move away slightly from the identity along a curve given by:
(t) = (
1
(t),
2
(t),
3
(t), )
For some parameter t such that
i
(t = 0) = 0. An innitesimal transformation near identity
along is given by Taylor Expansion for small t:
G(t) = G((t)) = I +
dG(t)
dt
|
t=0
t +O(t
2
) (1)
g

v()

v() =
dG(t)
dt
|
t=0

forms a Lie Algebra corresponding to the given Lie Group G.

Lie Algebra: Lie Algebra (g, [, ]) is a vector space over a eld F (say R or C) along with a
bracket operation [, ] dened by:
1. Bilinearity: [u +v, w] = [u, w] + [v, w] and [au, bv] = ab[u, v] a, b F, u, v, w g
2. Antisymmetry: [u, v] = [v, u] u, v g
3. Jacobi Identity: [u, [v, w]] + [v, [w, u]] + [w, [u, v]] = 0 u, v, w g
Innitesimal transformation near the Identity can also be written in terms of coordinate partials
using (Eqn. 1):
G(t) = I +
G(
i
)

i
|

i
=0
d
i
dt
|
t=0
t +O(t
2
)
also:

i
(t) =
i
(0) +
d
i
dt
|
t=0
t +O(t
2
) =
d
i
dt
|
t=0
t
G(t) = I +
G(
i
)

i
|

i
=0

i
(t) (2)
1
Change of basis transformation. D D

= SDS
1
for some S GL
n
(F)
3
Generators of Lie Group:
i
=
G(
i
)

i
|

i
=0
are called Generators of the Lie Group G.
In terms of these, innitesimal element is given by:
G(t) = I +
i

i
(t) (3)
Once generators in hand, one can easily reach the Lie Group by integrating many innitesimal
transformations together:
G(
i
) = lim
n
(I +
i

i
(t))
n
= e
n
i

i
(t)
= e

i
(4)
Remember
i
(t) are innitesimal parameters, while integrating innite of them will give me
nite parameters given by

i
= lim
n
n
i
(t)
{
i
} form a basis of lie algebra g. You can verify that they are linearly independent and form
a complete set. Therefore due to properties of Lie Algebra, [
i
,
j
] g.
[
i
,
j
] = f
ijk

k
(5)
f
ijk
are called Structure Constants of the Lie Algebra. Note that they are basis dependent; if
the parametrization of Lie Group is changed, generators of Lie Algebra will change, and hence
will change the structure constants. One can always choose a basis where structure constants
are completely antisymmetric
2
, i.e.:
f
ijk
=

f
jki
= f
kij
f
kji
= f
jik
= f
ikj
2
A fact that is not very trivial to digest. You can lookup any standard text in Lie Algebra for a detailed
proof, like in Chapter 2 of Ref. [5]
4
3 SU
2
Representations
If a system is invariant under certain transformation group, it is said to possess that particular
group symmetry. Whenever we have a physical quantum system with a group G of symmetries
acting on it, the n dimensional space of states H will carry a unitary representation
3
of G.
Why Unitary? Because U
n
s corresponding lie algebra is the set of all skew-hermitian matrices,
making the generators hermitian. Generators are thus Quantum Mechanical observables which
we would like them to be if they are conserved quantities
4
.
As we know laws of physics do not change under orientation preserving spatial rotations, all
physical systems possess rotational (or SO
3
) symmetry. But it turns out that the physical
systems carry representations of a larger isomorphic group SU
2
, where every element of SO
3
is
mapped to two elements in SU
2
. You must have already encountered this SU
2
symmetry in the
theory of Angular Momentum and Spin in Quantum Mechanics.
Hadrons
5
in Particle Physics carry an underlying SU
n
symmetry, basis behind which yet remains
unknown. It is this symmetry which leads to such nice geometric patterns in the arrangement
of Mesons and Baryons, and is cause of the success of Eightfold-Way Diagrams and the Quark
Model. With this motivation I will start with unitary representations of the simplest group
among SU
n
, i.e. SU
2
, relating it time to time with familiar structure of Angular Momentum.
At the end, I will establish their importance in the Baryons and Mesons structure.
3.1 spin j representation of SU
2
For all positive integers n, there exists an n dimensional unitary representation of SU
2
called
spin j (n = 2j + 1), given by:
spin j : SU
2
U
2j+1
j =

0,
1
2
, 1,
3
2

(6)
The nomenclature is derived from Quantum Mechanics, but has more general sense to it. spin j
depicts the representation of SU
2
for 2j + 1 dimensional state space H. Let me pick up three
appropriate
6
generators {J
j
1
, J
j
2
, J
j
3
} for corresponding spin j representation. Commutator re-
lationships between them are given by:
[J
j
l
, J
j
m
] =
lmn
J
j
n
where,
ijk
is the fully antisymmetric Levi-Cevita Tensor, given by:

ijk
=

1 for cyclic permutations of 1,2,3

1 for anti-cyclic permutations of 1,2,3
0 if any two indices are same
Since none of the generators commute, I can diagonalize any one of them at a time, which is
conventionally chosen to be J
j
3
. Thus basis states of H are eigenvectors of J
j
3
. Action of J
j
1
and J
j
2
will give some non trivial combinations of basis states, which are not much of interest.
What are of interest are step-up and step-down operators, given by:
J
j

=
1

2
(J
j
1
J
j
2
) (7)
3
Representation in the unitary group U
n
.
4
By Noethers Theorem
5
Baryons and Mesons are collectively termed as Hadrons.
6
Such that the structure constants are completely antisymmetric.
5
Note that:
[J
j
3
, J
j

] = J
j

[J
j
+
, J
j

] = J
j
3
If I consider a basis state | with eigenvalue of J
j
3
as m:
J
j
3
| = m|
J
j
3
(J
j

|) = (m1)J
j

|
Thus, step-up and step-down operators alters the eigenvalues of J
j
3
by 1. Basis states of spin j
can be labelled
7
with the eigenvalues of J
j
3
(denoted by m
j
) and j:
|j, m
j
m
j
= {j, j 1, , j} (8)
These are commonly called Spin (2j + 1)-tuple. If we have any one basis state, all others can
be reached by J
j

operators.
3.1.1 Fundamental Representation of SU
2
Fundamental Representation of SU
2
is the map from SU
2
to SU
2
itself, i.e. spin
1
2
. Generators
for spin
1
2
are given in terms of well known Pauli Matrices:
J
1
2
i
=
1
2

i
(9)

1
=

0 1
1 0

;
2
=

0
0

;
3
=

1 0
0 1

And, step-up and step-down operators are:

J
1
2
+
=

0 1
0 0

; J
1
2

0 0
1 0

As spin
1
2
representation contains full information about SU
2
, all the higher dimensional rep-
resentations can be reached from it using Tensor Product.
3.2 Tensor Product of Representations
This is the generalization of what is called the Addition of Angular Momentum in Quantum
Mechanics. If I have a composite system of two independent components both carrying SU
2
representations spin j
A
and spin j
B
respectively, its states are denoted by:
|j
A
, j
B
, m
j
A
, m
j
B
= |j
A
, m
j
A
|j
B
, m
j
B

We have total (2j

A
+ 1)(2j
B
+ 1) such states. A theorem in Representation Theory
8
says that
the tensor product of two SU
2
representations spin j
A
and spin j
B
can be decomposed into the
direct sum of irreducible representations spin j given by:
j = |j
A
+j
B
|, |j
A
+j
B
1|, , |j
A
j
B
| (10)
m
j
= m
j
A
+m
j
B
(11)
7
I have skipped the explicit proofs. See Chapter 3 of Ref. [5] for details.
8
See section 9.4.1 of Ref. [6] for details.
6
With respective basis states:
|j
A
, j
B
, j, m
j
=

m
j
A
+m
j
B
=m
j
C(m
j
A
, m
j
B
; j, m
j
) |j
A
, j
B
, m
j
A
, m
j
B
(12)
m
j
is the eigenvalue of J
j
3
operator for corresponding representation. Cs are called Clebsch-
Gordan coecients tabulated in Review of Particle Physics (1982), numerical values of whose
does not bother us much at this point.
There is a nice diagrammatic way to handle these representations which might not be so illu-
minating at the moment, but it nds its importance in higher groups SU
n
, and give base for
Eightfold-Way Diagrams in Particle Physics. spin j representation is diagrammatically repre-
sented on an axis of m
j
values, with base states marked on it. J
j

takes one state to another

by step-up and step-down operation.
Figure 1: Diagrammatic Representation of spin j
Now consider (spin j
A
spin j
B
). The tensor product is generally written in symbolic form as
follows:
(2j
A
+ 1) (2j
B
+ 1) = (2|j
A
+j
B
| + 1) (2|j
A
j
B
| + 1) (13)
Draw one of the representation (say j
A
). Now draw the other representation (say j
B
) on each of
the states of the rst, overlapping it with m
j
B
= 0 (See Fig 2). Forget the j
A
representation; for
a corresponding value of m
j
you get a multiplet of j
B
states. Take a state from each available
multiplet, and form the highest j representation, spin |j
A
+ j
B
|. Respectively from residual
states, take one state each and form the second highest representation, and so on until you run
out of states. You will see that the state decomposes as expected and will furnish you 2j
B
+ 1
representations.
States with some particular value of m
j
in the resultant representations are actually linear
combinations of states from parent representations with m
j
A
+m
j
B
= m
j
, coecients of which
are given by Clebsch-Gordan coecients (see Eqn. 12).
3.3 Isospin
The theory developed above looks similar to the theory of Angular Momentum in Quantum
Mechanics. Reason being that the theory of Angular Momentum is the result of rotational SU
2
symmetry in our system. There are also other systems that possess SU
2
symmetry, for instance
the theory of Isospin in Nuclear Physics.
Similarity in the masses of Neutrons and Protons and their coexistence in nuclei led people to
believe that they are merely dierent states of the same particle- nucleon, more or less the same
way electrons have two spin states- and . This new property was termed Isospin, and was
believed to manifest the charge dierence between the two. I will however refrain from the topic
of charge, until we reach Quark Model.
7
Figure 2: Tensoring of Representations
Isospin is also considered to be a consequence of an underlying SU
2
symmetry same as spin.
Nucleons carry (I =
1
2
) fundamental representation of Isospin group, with I
3
values
1
2
(proton)
and
1
2
(neutron). Protons and Neutrons thus form a Isospin Doublet.
There are other hadrons which carry Isospin symmetry. Some of them are:
Isospin Baryons Mesons
Singlet (),(

) (),(

)
Doublets (n,p
+
),(

,
+
) (K
0
,K
+
),(K

,K
0
)
Triplets (

,
0
,
+
) (

,
0
,
+
)
Quadruples (

,
0
,
+
,
2+
)
3.3.1 Bound States of Nucleons
Nucleons possess the fundamental representation (spin
1
2
) of isospin, thus all higher representa-
tions can be reached from their composite, commonly known as nuclei. Tensoring two nucleons
gives two irreducible representations I = 1, 0: 2 2 = 3 1.
Figure 3: Two nucleon bound states.
8
I = 0 (p-n antisymmetric state) is the ground state of two nucleon system, commonly known
as Deuterium. I = 1 are the excited states, which includes p-p (di-proton), n-n (di-neutron)
and n-p symmetric (excited deuterium) states.
Similarly all the higher nucleon bound states can be constructed from the symmetry argument,
stability or instability of whose vests on energy considerations. I have given below the three
nucleon bound states, most of which are not found in nature.
2 2 2 = (3 1) 2 = 4 2 2
Figure 4: 3 2 and 1 2 diagrams for three nucleon bound states. Angular brackets signify the
linear combinations of the arguments and its permutations.
By calculating the Clebsch-Gordan coecients it can be veried that no two states are the
same, though they appear to be at rst. nnp are the states of Tritium (
3
H
1
) and npp are
the states of a Helium Isotope (
3
He
2
); (ppp) and (nnn) are non physical.
9
4 SU
3
Representations
Next we move to SU
3
symmetries. In the current context, they nd their prominent application
in Particle Physics in form of the Eightfold-Way diagrams. The representation theory for SU
3
can be developed more or less in the same way we did for SU
2
, but the mathematical jugglery
involved gets much complicated. I will try to keep the things fairly simple and understandable
on the cost of mathematical preciseness, for which you must spare me. You can nd a detailed
treatment of lie algebra representation theory in Ref [5], where most of the loopholes I will be
leaving in my discussion are covered.
SU
3
has dimension n
2
1 = 8, thus its lie algebra su
3
(33 traceless hermitian matrices) have 8
generators. I will denote them by J
i
. Commutation relations between the generators are given
by:
[J
i
, J
j
] = f
ijk
J
k
(14)
Recall that the structure constants f
ijk
are completely antisymmetric, and its non vanishing
components (along with their permutations) are given by:
f
123
= 1; f
458
= f
678
=

3
2
f
147
= f
165
= f
246
= f
257
= f
345
= f
376
=
1
2
With these relations one can easily verify that out of all generators only J
3
and J
8
commute.
This is not surprising because we have two independent 3 3 traceless diagonal matrices, and
we choose to diagonalize J
3
and J
8
. Similar to step-up and step-down operators in SU
2
, here
we have:
T

= J
1
J
2
; T
3
= J
3
V

= J
4
J
5
; V
3
=

3
2
J
8
+
1
2
J
3
U

= J
6
J
7
; U
3
=

3
2
J
8

1
2
J
3
Surprisingly they form three sets of SU
2
sub-algebras:
[T
3
, T

] = T

; [T
+
, T

] = T
3
[V
3
, V

] = V

; [V
+
, V

] = V
3
[U
3
, U

] = U

; [U
+
, U

] = U
3
But not all three of them are independent, as U
3
and V
3
are linear combinations of T
3
and J
8
.
T
3
, V
3
and U
3
are observables with eigenvalues denoted by t, v and u respectively, which are
related by:
t = v u (15)
These values can be raised and lowered by T

, V

and U

operators. Due to this interdepen-

dence of observables, there is a degeneracy in our system, i.e. we have dierent states with
same values of t, v and u.
Same as SU
2
we can give an expression for the dimension of representation in terms of maximal
eigenvalues; which I will state here without a proof. If I choose to pick the base state with
maximal eigenvalue v = j
v
, dimension of representation is given by:
n = (|j
v
| + 1)(2|t| + 1)(2|u| + 1) = (|j
v
| + 1)(2|t| + 1)(2|j
v
t| + 1) (16)
where v and u are other two quantum numbers for that particular state.
10
4.1 Diagrammatic Representation
I will skip the math, but what is of more importance to us is the diagrammatic illustrations of
the SU
3
representations and tricks to handle them. Extending our ideas developed for SU
2
, we
add another dimension to our diagrams as we have two independent observables this time. We
need to depict 3 SU
2
sub-algebras in a 2-D diagram, so a triangle will be the most obvious and
suitable choice.
Figure 5: D(0, 4) or D(15): Non-degenerate 15 Dimensional SU
3
Representation.
In Fig. 5, I have a 15 Dimensional Representation, with step-up and step-down operators indi-
cated. This is actually one of the simplest cases because for v = j
v
, |t| = 0; such representations
(and also |t| = j
v
) are called non-degenerate representations, as they show no degeneracy. In
general |t| can take any value from 0 to j
v
, and we can have degeneracy in our system. I have
shown in Fig. 6, a 42 dimensional degenerate representation with j
v
=
5
2
and t = 1. Note the
shape which is a semi-regular hexagon, with two set of equal side lengths.
There is a more natural way to denote these diagrams other than j
v
and t. Dene:
= 2|t|
= 2|j
v
t|
n = ( + 1)( + 1)
( + + 2)
2
(17)
and are nothing but the top-most and bottom-most sides of the semi-regular hexagon
respectively. The representation is called D(, ) or D(n) (if > ) or D(n) (if < ).
11
Figure 6: D(2, 3) or D(42): Degenerate 42 Dimensional SU
3
Representation.
Degeneracy is depicted by multiplets at the particular position in diagrams. There is an easy
algorithm to plot it:
1. Plot your representation without degeneracy.
2. Mark states of outermost semi-regular hexagon layer as non-degenerate. Next inner semi-
regular hexagon layer is doubly degenerate and so on, until you reach a triangle.
3. States inside the triangle will have same degeneracy as the periphery of triangle.
Figure 7: D(5, 2) or D(81): Degeneracy in diagrammatic representation.
12
4.2 Fundamental Representation of SU
3
Fundamental representation(s) of SU
3
are the map to SU
3
itself. There are two fundamental
representations in this case: D(1, 0) (or D(3)) and D(0, 1) (or D(3))
Figure 8: D(3) and D(3): Fundamental Representations of SU
3
.
Generators of the fundamental representations are given by Gell-mann matrices:
J
i
=
1
2

i
(18)

1
=

0 1 0
1 0 0
0 0 0

;
2
=

0 0
0 0
0 0 0

;
3
=

1 0 0
0 1 0
0 0 0

4
=

0 0 1
0 0 0
1 0 0

;
5
=

0 0
0 0 0
0 0

6
=

0 0 0
0 0 1
0 1 0

;
7
=

0 0 0
0 0
0 0

;
8
=

1 0 0
0 1 0
0 0 2

Readers might want to check the corresponding commutation relations. Same as SU

2
, all the
further representations can be reached by tensoring these fundamental representations.
4.3 Tensor Product of SU
3
Representations
Method we adopt for tensor product is again diagrammatic. Similarly like SU
2
, overlap one
of the diagrams on all the states of the other diagram (Caution! You do not have another
dimension at your ease to do that). Look at the outermost outline of the resultant gure
you get, a hexagon, sides of which will give the corresponding values of and . Plot the
corresponding diagram, and omit the used states from your overlapped gure. Now again look
at the outermost outline after omission, new values of and will give you another diagram.
Continue doing this until you are out of states; these are the resultant diagrams after tensoring.
In Fig. 9 I have depicted following tensor product:
D(2, 1) D(0, 2) = D(2, 3) D(3, 1) D(1, 2) D(2, 0) D(1, 0)
15 6 = 42 24 15 6 3
Similar steps can be adopted to reach all SU
3
representations from the fundamental represen-
tations D(3) and D(3).
13
Figure 9: Tensoring of Representations. D(2, 1) D(0, 2) gives D(2, 3) D(3, 1) D(1, 2)
D(2, 0) D(1, 0)
14
5 SU
4
and Higher Representations
5.1 SU
4
Representations
We will not be discussing SU
4
representations mathematically, but it is not hard to guess what
we expect based on our earlier experience. There should be 4
2
1 = 15 generators, out of which
3 will be diagonal (we can have 3 independent 4 4 traceless matrices). Rest 12 non-diagonal
generators will give rise to 6 step-up and 6 step-down operators. Hence we will get 6 SU
2
sub-
algebras (denoted by T, U, V, W, X, Y with their eigenvalues t, u, v, w, x, y), out of which only
3 will be independent. Again due to commutating generators we will have degeneracy in the
system.
Figure 10: Diagrammatic Representations of SU
4
- Tetrahedron, Octahedron and 14-hedron.
Inner details have been skipped for visual simplicity.
Non-degenerate representations diagrammatically take form of a regular tetrahedron in 3-D,
which has 6 edges (corresponding to SU
2
sub-algebras) and 4 vertices. Degenerate representa-
tions can take form of semi-regular octahedron and 14-hedron.
Fundamental representations are the map from SU
4
to SU
4
itself, diagrammatically which are
nothing but regular unit-sided tetrahedrons with a base-state at each vertex.
Figure 11: Fundamental Representations of SU
4
Same as earlier cases, all representations of SU
4
can be reached from fundamental represen-
tations. If your 3-D visualization is strong, you might be able to tensor these fundamental
representations to get structures given in Fig. 10.
5.2 Higher SU
n
Representations
Even diagrammatic representations of SU
n
for n > 4 are hard to tackle, because diagrams are
n 1 dimensional, and at best we can comprehend 3-D objects. Though the theory remains
similar it is of less use, and we have to go back to rigorous group theoretical formalism which
15
I will not be discussing here. We can still get some information about our SU
n
system by
studying its SU
3
sub-algebras and hence projecting it to our familiar 3-D.
16
6 Quark Model
Now when we have assembled all the required mathematical tools, we are ready to discuss the
Physics part this paper is supposed to cover. There is an apparent
9
global SU
n
symmetry
in particle physics called Flavour Symmetry, much similar to the rotational symmetry in
Quantum Mechanics, which governs the pattern Baryons and Mesons families exhibit. But
unlike rotations, symmetry involved here is more of a hypothesis posed by the Quark Model ,
which best explains the similarity in hadrons found in nature.
By the mid of 19th Century the Hadron zoo had grown into a complete mess. Much like the
elements in the periodic table, family of hadrons was growing huge, and was awaiting their
own periodic table. Finally in 1961, Gell-mann introduced his Eightfold-way Diagrams, and
arranged Hadrons in weird gures based on their charge and strangeness. Though the gures
might appear completely empirical to someone taking his rst course in Particle Physics, they
appear naturally from Group Theory and are nothing but the SU
3
representation diagrams.
This gave a strong sign of an underlying SU
3
symmetry in the system, which was soon realized
by Gell-mann and Zweig, who independently gave the Quark Model for Hadrons in 1964.
Inspired from nucleons which were understood as Isospin Doublet, fundamental representation
of SU
2
which generated all the nuclei occurring in nature; it was proposed that there exists
a fundamental SU
3
Triplet termed Quarks, which generated all the Hadrons. The feasibility
or limitations of the model can be looked-up in any standard text on Particle Physics, I will
concentrate here on how does it explain the Hadron structure. On the course we shall not
necessarily trace the historical development, and will assume all quarks we need are known. We
shall inherit following facts from theory:
1. Antiparticles have all the quantum numbers with sign reversed compared to their corre-
sponding particles.
2. Only possible quark combinations are qq (mesons and anti-mesons), qqq (baryons) and
qqq (anti-baryons), where q is the antiparticle of q.
3. u, c, t have charge
2
3
, while d, s, b have charge
1
3
in electronic units.
6.1 2 Quark Model - Up (u) and Down (d)
Although the original idea was proposed with 3 quarks, it is worth looking at 2 quark system u
and d to start with. Not only it is simpler as it exhibits SU
2
symmetry, it will give us an insight
on how can we generalize our theory to 4 and higher quark systems. Initially known hadrons
(non-strange) were just composed of u and d quarks. Inspired from nuclear systems, u and
d can be considered as an Isospin Doublet
10
with I
3
values
1
2
respectively. Correspondingly
u and d form a Isospin Doublet with I
3
=
1
2
respectively. These will serve as fundamental
representations for this model.
Figure 12: u and d quarks. Fundamental representation of 2 quark model.
9
The symmetry is apparent/approximate and not exact. I shall come to this point towards the end.
10
Right now we have no reason to believe that this Isospin is the same Isospin we encountered in nuclear case,
but as we will see soon it is.
17
Charge in this model is nothing but a mere shift of origin:
q = I
3
+
1
6
for quarks; q = I
3

1
6
for antiquarks
We can dene quantities like Up-ness and Down-ness to measure the content of u and d quarks
in our composite system.
U = I
3

1
2
; D = I
3

1
2
(+ for quarks, - for antiquarks)
Positive value of U and D tells the number of u and d quarks in the system, while negative
values signify the presence of corresponding anti-particles. Similar to isospin, U, D and q also
adds up in composite systems.
Mesons: As we know Mesons are qq bound states, they can be formed by q q.
2 2 = 3 1
Figure 13: u and d quark Meson states.
We get an Isospin triplet (

,
0
,
+
) and an Isospin singlet (
0
), all of which are observed in
nature.
Baryons: Baryons are one step complicated than mesons as they are qqq bound states. They
can be reached by q q q.
2 2 2 = (3 1) 2 = 4 2 2
Figure 14: u and d quark Baryon states. Angular Brackets signify linear combinations of the
arguments and its permutations.
18
We get an Isospin quadruplet (

,
0
,
+
,
2+
) and two linear combinations of Isospin dou-
blet (n
0
,p
+
). The Isospin Doublet is what we see commonly around us, which provides the
fundamental representation for all nuclei (see Sec. 3.3.1).
Similarly one can form 2 quark representation of anti-baryons:
Figure 15: u and d quark anti-baryon states.
6.2 3 Quark Model - Up (u), Down (d) and Strange (s)
Now we are ready to add another quark to the game, s or strange quark. In the beginning of the
hunt for new elementary particles, some strange (so were they named) hadrons were observed,
which could be explained by inclusion of another particle to our set of quarks and going to
SU
3
instead of SU
2
. u, d, s form the fundamental representation for this model, out of which
sub-algebra u-d reduces to our SU
2
Isospin algebra.
Figure 16: D(3) and D(3): Fundamental Representations of SU
3
.
Charge in this model is dened to be
q =
2
3
(t +v)
Whereas U and D are revised to be:
U =
2
3
(t +v)
1
3
; D =
2
3
(u t)
1
3
(+ for quarks, - for antiquarks)
Similar to these we dene a new quantity Strangeness:
S =
2
3
(v u)
1
3
(+ for quarks, - for antiquarks)
+ve value of S tells the number of s quarks in our composite system, while -ve value tells the
number of s anti-quark
11
. Putting S = 0 we will land back in u-d sub-algebra, and above
formulas for U, D and q will reduce to their earlier SU
2
forms.
Mesons: Now it is an easy job to reach Mesons (qq); just nd 3 3.
3 3 = 8 1
11
Caution! In Physics literature generally S is termed as strangeness. It is just a matter of historical
convention.
19
Figure 17: u, d and s quark Meson states.
They are generally called Meson Octet and Singlet. What we have just derived is nothing but
Gell-manns rst Eightfold-Way diagram for Mesons. It is more conventional to draw it on q
and (S) axes as follows:
Figure 18: Meson Octet in Eightfold-Way representation.
Baryons: Baryon states (qqq) in 3 quark Model are given by (See Fig. 19):
3 3 3 = (6 3) 3 = 10 8 8 1
They take form of Baryon Decuplet, Octet and Singlet. These are Eightfold-Way diagrams for
Baryons, which can also be expressed in terms of q and S (Fig. 20).
Figure 20: Baryon Decuplet and Octet in Eightfold-Way representation.
20
Figure 19: u, d and s quark Baryon states. 3 3 = 6 3; 6 3 = 10 8; 3 3 = 8 1.
Symmetric linear combination of the two Octets formed after tensoring gives the Gell-manns
Baryon Octet, while the singlet is not observed in nature.
Similarly anti-baryons (qqq) in 3 quark model can be reached by:
3 3 3 = (6 3) 3 = 10 8 8 1
Figure 21: Anti-baryon Decuplet and Octet in Eightfold-Way representation.
21
6.3 4 Quark Model - Up (u), Down (d), Strange (s) and Charm (c)
In 1974, a new quark c or Charm was discovered
12
. Soon after its discovery, whole new range
of hadrons having charm content were reported. Quark Model had to be modied just slightly
to include this new member. Same as we went from SU
2
to SU
3
by adding s quark, inclusion
of c modies the underlying symmetry to SU
4
. (u, d, s, c) and (u, d, s, c) form the fundamental
representations for this model:
Figure 22: Fundamental Representation for 4 quark model.
Charge in this model will be given by:
q =
1
2
(t +v +w +x)
1
6
(+ for quarks, - for antiquarks)
Whereas up-ness, down-ness, strangeness and charmness are given by:
U =
1
2
(t +v y)
1
4
; D =
1
2
(u t w)
1
4
(+ for quarks, - for antiquarks)
C =
1
2
(w +y +x)
1
4
; S =
1
2
(u v x)
1
4
(+ for quarks, - for antiquarks)
Mesons: Mesons, i.e. bound states of qq take form of 14-hedron (15-tuplet) in this model. I
will not depict the overlapping because it is too messy but the tensoring goes as:
4 4 = 15 1
Figure 23: 4 quark meson states. Meson 15-tuplet and singlet.
12
Not exactly c was discovered, but its bound state cc in form of J/ meson.
22
Ground states of all observed u, d, s and c Mesons arrange themselves nicely in the above 3-D
structure.
Baryons: Baryons states in 4-quark model form 20-tuplets (tetrahedron and octahedron) and
quadruplet (tetrahedron). Though quadruplet is not found in nature.
(4 4) 4 = (10 6) 4 = 20 20 20 4
Figure 24: 4 quark Baryon states. Baryon 20-tuplets and quadruplet. Quadruplet is however
not found in nature.
Similarly, anti-baryon diagrams are nothing but a inversion about origin of these diagrams.
6.4 Complete Quark Model
Near 1980s nal two quarks, top (t) and bottom (b) were added to the family, rising the total
number to 6. Till now it should be obvious that the corresponding symmetry will rise to SU
6
.
As I have mentioned earlier, for any n > 4 diagrammatic representation of SU
n
is not much
fruitful, so Eightfold-Way Diagrams meet a limitation. But if we have some insight of higher
dimensions, we can project out SU
4
sub-algebras out of SU
6
and get some information about
the system; after all at best we have only 3 quark bound states. But this procedure is not much
worth either, because t and b quarks are too heavy and rarely form any bound state.
23
7 Summary
I will recapitulate what have we accomplished in more general notation. SU
n
, set of all n n
unit determinant unitary matrices, form a Lie Group. If a physical system is carrying SU
n
symmetry, its state space H will carry a unitary representation of SU
n
. For example, quantum
mechanical systems possess SU
2
symmetry corresponding to Angular Momentum and Spin, and
thus carry a 2j + 1 dimensional unitary representation of SU
2
. Here j is a quantum number
that corresponds to angular momentum state or spin of the system.
Similarly Hadrons in Particle Physics possess SU
6
avour symmetry. Basis states of the funda-
mental representations corresponds to the six quarks u, d, s, c, t, b and six antiquarks u, d, s, c, t, b.
Composite systems of these 2 fundamental representations gives rise to all Mesons (qq) and
Baryons (qqq,qqq). SU
3
SU
6
sub-algebra corresponding to quarks u, d, s gives rise to the
Gell-manns Eightfold way diagrams which provided basis for the Quark Model of hadrons.
Similarly SU
4
SU
6
sub-algebra corresponding to quarks u, d, s, c gives rise to the pyramidal
3-D Eightfold-Way diagrams.
The nal impression I want to make here is that the complete theory of Hadron Structure (or
avour symmetry) can be developed just from the argument of symmetry and group theory.
This exhibits a strong correlation between Physical Systems and Mathematics, even stronger
examples of which can be found scattered all over in Physics.
8 Discussion
Concept of Symmetry is not new to physicists. It is the physics analogue to the word Beautiful.
Ellipticity of the bound orbits, or the conservation of Energy-Momentum, or the spherical struc-
ture of cosmic bodies, all are consequences of some underlying symmetry. The same argument
of symmetry led people to believe in existence of a Theory of Everything and Super-symmetry.
But it sometimes makes one wonder - why is the world so symmetric or ordered, it could have
been a lot more messier. A pure mathematical topic Representation Theory of Unitary Matrices
and a pure physical theory of Hadrons in Particle Physics - the correspondence between the two
is both surprising and elegant.
We have certain cases of exact SU
n
symmetries in Physics, i.e. state vectors can not be distin-
guished with each other apart from the generators eigenvalues, for example Gauge Theories and
Lorentz Transformations. But the symmetry we have at hand, avour symmetry of Hadrons
is an approximate one. Reason being the mass of quarks which is way dierent from each
other, and one can easily distinguish between quarks on basis of their masses. Apart from the
masses too, quarks come in 2 varieties - charge 2/3 and 1/3, which again makes the symmetry
approximate.
Even if approximate, symmetry in Hadrons is of much theoretical importance. After all even
periodic table of elements is an approximate symmetry, but a whole branch of chemistry is
dedicated to its study, and no one can argue upon its importance. Eightfold-Way Diagrams are
also important in their own right; they classify hadrons in a coherent fashion and help us study
their properties and structure.
My discussions in this term paper are mainly directed towards the Eightfold-Way Diagrams and
avour symmetry, but in a similar fashion one can develop a theory for any symmetry whatso-
ever. In the discussion we had in Section 3.2, about SU
2
symmetries and their representations,
just change of the terminology will fetch us the familiar quantum mechanical theory of angular
momentum and spin. Local Symmetries (or Gauge Symmetries), which I have not touched at
24
all here, can also be developed on the same grounds except that the transformations will now be
space-time dependent. This will demand stronger symmetry in the system, and we will have
stricter conservation laws. For instance theories of fundamental forces - Electromagnetism (U
1
),
Weak Interactions (SU
2
) and Strong Interactions (SU
3
) which comprises the Standard Model,
have Gauge Symmetries. All of them have their corresponding charge which is conserved locally
at every space-time point.
References
[1] David Griths, Introduction to Elementary Particle Physics.
[2] Francis Halzen, Alan D. Martin, Quarks and Leptons.
[3] S. D. Joglekar, Mathematical Physics - Advanced Topics.
[4] Krishna Kaipa, Lecture Notes on Matrix Groups, Lie Algebras, SU
2
and SO
3
.
[5] Howard Georgi, Lie Algebra in Particle Physics.
[6] Peter Woit, Quantum Mechanics for Mathematicians: Fall 2012 Course Notes.
[7] Dummit Foote, Abstract Algebra.
[8] Arfken, Mathematical Methods for Physicists.
25