You are on page 1of 38

M ASTERS T HESIS

Cluster Algebraic structures in


Supersymmetry

Author: Supervisor:
Zain Ateeq Dr. Rizwan Khalid

A thesis submitted in fulfillment of the requirements


for the degree of Master in Physics

Department of Physics
L AHORE U NIVERSITY OF M ANAGEMENT S CIENCES
iii

Acknowledgements
****Not finalized yet****
v

LAHORE UNIVERSITY OF MANAGEMENT SCIENCES

Abstract
Department of Physics

Master in Physics

Cluster Algebraic structures in Supersymmetry


by Zain Ateeq

****Not finalized yet****


vii

Contents

Acknowledgements iii

Abstract v

1 Introduction 1

2 Cluster Algebras 3
2.1 Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1 Classification of Graphs . . . . . . . . . . . . . . . . . . . 3
2.1.2 Representing Graphs . . . . . . . . . . . . . . . . . . . . . 4
2.2 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.1 Seed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.2 Mutation . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Quiver Mutation . . . . . . . . . . . . . . . . . . . . . . . 5
Seed Mutation . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Cluster Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.1 A2 Cluster Algebra . . . . . . . . . . . . . . . . . . . . . . 8

3 Spinors 13
3.1 Dirac Equation and the Dirac Spinor . . . . . . . . . . . . . . . . . 13
3.2 Weyl Spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3 Charge Conjugation . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.4 Majorana Spinors . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 Lorentz Invariants and Transformations 19


4.1 Lorentz Transformation of Spinors . . . . . . . . . . . . . . . . . . 19
4.2 Lorentz Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3 Supersymmetric Transformations of Fields . . . . . . . . . . . . . . 22

5 Generators of Supersymmtery 23
5.1 Generators and Their Algebra . . . . . . . . . . . . . . . . . . . . 23
5.1.1 General Discussion . . . . . . . . . . . . . . . . . . . . . . 23
5.1.2 Generators and Algebra of Supersymmetry . . . . . . . . . 25
5.2 Nonclosure SUSY Algebra and the Need of an Auxilary Field . . . 27
ix

List of Figures

2.1 Example of a graph. . . . . . . . . . . . . . . . . . . . . . . . . . . 3


2.2 Example of an undirected graph. . . . . . . . . . . . . . . . . . . . 4
2.3 Example of a Seed. . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4 Example of a graph. . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5 Example of a graph. . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.6 Example of a graph. . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.7 Example of a graph. . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.8 Example of a graph. . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.9 Exchange graph of A2 algebra. . . . . . . . . . . . . . . . . . . . . 12
1

Chapter 1

Introduction

Not finalized yet.


3

Chapter 2

Cluster Algebras

This chapter is an introduction to cluster algebras, starting from the basic definitions
to the algebras which are used in supersymmetry. The details of how cluster algebras
are applied in supersymmetry are not discussed here and will be presented in the later
chapters.

Cluster algebras which are commutative algebras were first conceived by two
mathematicians S. Fomin and A. Zelevinsky and are defined by operation of iterated
mutations in different combinations. [1]*** To understand cluster algebras, it is
imperative to first understand some basic definitions and terminologies. This section
introduces the fundamentals which are at the least required to grasp the idea of cluster
algebras.

2.1 Graphs
A graph in simplest terms is a collection of nodes and edges. It is an ordered pair
G = (V, E) in which V is a set of vertices and E is the set containing the edges. A
simple example is if we have V = (a, b, c, d) and E = ((a, c), (b, c), (a, d), (d, c)),
then we can make the graph shown in Figure 2.1.

2.1.1 Classification of Graphs


There are two types of graphs, undirected graphs and directed graphs. An undirected
graph is a graph with its edges having no direction like the graph shown in Figure
2.1. A directed graph however, has a direction to each of its nodes as shown in Figure
2.2.

2 3 4

F IGURE 2.1: Example of a graph.


4 Chapter 2. Cluster Algebras

2 3 4

F IGURE 2.2: Example of an undirected graph.

2.1.2 Representing Graphs


There are three data structures by which one can represent a graph, which are matrix,
list and set. Using the matrix representation, the graph in Figure 2.1 is written as

 
0 0 1 1
0 0 1 0
G=
1
. (2.1)
1 0 1
1 0 1 0

And that of Figure 2.2 is represented as


 
0 0 1 −1
0 0 1 0
G= −1 −1 0
. (2.2)
1
1 0 −1 0

Each element, say Gij , in the matrix represents an edge from node i to node j.
In case of undirected graphs, the matrix is always symmetric because an edge from
node i to j will also be an edge from node j to i. But directed graphs are skew
symmetric because of the fact that an edge from i to j will equal ******. The size
of matrix depends on the number of nodes and it is always a square matrix.

2.2 Definitions
Cluster Algebras appear differently on different configurations of initial seed. A seed
upon mutation, which we will see shortly is an operation, gives rise to a new seed
which can be mutated again to form more seeds. This repetitive mutation if somehow
produces the same initial seed then that means that it is periodic in mutations and its
algebra is called cluster algebra of finite type. But if there is no periodicity in seeds
after mutation then the algebra is called cluster algebra of infinite type.

2.2.1 Seed
A seed is defined by two things, quivers and their cluster variables. Quiver is a
directed graph with one condition that it must not contain self loops i.e an arrow
from a node to itself. Just like a directed graph, we can represent a quiver by a skew
symmetric square matrix. Figure 2.2 can be taken as a quiver Q and equation 2.12 as
its quiver matrix representation.
2.2. Definitions 5

2 3 4

F IGURE 2.3: Example of a Seed.

A quiver with each of its nodes assigned an independent variable becomes a seed.
These variables are called cluster variables and the set containing all those variables
is called a cluster. For our quiver example in Figure 2.2, we assign a1 , a2 , a3 and
a4 to its nodes as the cluster variables and it forms a seed as shown in Figure 2.3.
Equation below shows cluster variables as a cluster a of this seed:

a = {a1 , a2 , a3 , a4 }. ∗ ∗∗ (2.3)

Thus a seed S is an ordered pair S = (a, Q) where a is a cluster and Q is a quiver.

2.2.2 Mutation
Mutation is an operation which when applied on a seed results in a new seed. This
operation transforms the quiver differently based on which vertex it was applied. Not
only the quiver, but the cluster variables are also changed upon every mutation.

Quiver Mutation

Quiver mutation k is defined as a map µk (Q) = Q . This process is applied on a
vertex of a quiver by following rules:

1. Invert the direction of all the arrows originating or terminating at the target
vertex.

2. For paths which are traversing node k in the middle from a node, say i, to a
node, say j, add a new arrow from i to j.

3. The last step is to remove all the two cycles which were produced in the second
step.

Take the quiver Q in Figure 2.2 and apply mutation on vertex 3. The mutation is

represented by µ3 and the resultant quiver Q in terms of Q can be written as:

Q = µ3 (Q). (2.4)

Now to follow the steps to mutate the quiver, first invert the direction of arrows
as said in rule one. New changes are shown by red arrows and the mutating vertex is
filled with blue. We get the following graph:
6 Chapter 2. Cluster Algebras

2 3 4

After introducing new arrows as described in rule two, the output graph becomes

2 3 4

In the last step, we have to remove all the two cycles and we get

2 3 4

In the second step, multiplicity is taken into account, i.e. if there are p arrows
from i to k and q arrows from k to j, then the number of new added arrows equals to
the product p.q.

Final quiver Q can easily be written in its matrix form by looking at the number
of arrows between each nodes. Another way of getting the matrix is by using the
following relations
(
−Qij if k ∈ {i, j}
Q′ij = |Qik |Qkj +Qij |Qkj |
Qij + 2
otherwise
(2.5)

−Qij
 if k ∈ {i, j}
= Qij + Bik |Qkj | if sgn Qik = sgn Qkj

Qik otherwise.

Whatever method is used the matrix for the mutated quiver becomes

 
0 0 −1 0
′ 0 0 −1 1 
Q = . (2.6)
1 1 0 −1
0 −1 1 0
2.2. Definitions 7

F IGURE 2.4: Example of a graph.

Seed Mutation
Seed mutation is a map from seed to seed defined by the relation
′ ′ ′
S = µk (S) = µk (a, Q) = (a , Q ). (2.7)
It involves both mutation of a quiver and mutation of its cluster variables. We
know how quiver transformed upon mutation µ3 from previous example. The cluster
variables are mutated following the relation
!
1 Y Y
a′k = ai + aj . (2.8)
ak i→k k→j

What is happening here is basically the summation of two things. The first is
the product of all the cluster variables pointing towards the mutating node k and the
second is the product of all the cluster variables outgoing from k. Lets apply this to
our seed shown in Figure 2.3
Now upon mutation µ3 , this seed takes the form given in Figure 2.4. Keep in mind
that it is the same quiver as the one obtained in the previous example after quiver
mutation but with a bit of rearrangements and its labels replaced with its cluster

variables to form a seed. The cluster variable a3 is transformed to a3 which after
using equation 2.7 follows the relation
′ 1
a3 = (a1 a2 + a4 ). (2.9)
a3
If a seed is mutated twice with the same mutation, it results in the same initial
seed. It means that if we apply mutation µ3 on seed in Figure 2.4, then the result will

be the initial seed shown in Figure 2.3. Figure 2.5 shows the seed mutation µ3 (S )

going back to S and equation 2.10 shows how a3 when mutated again, gives a3 back,
thus reproducing the initial quiver.

′′ 1
a3 = ′ (a4 + a1 a2 ),
a3
′′ a3
a3 = (a4 + a1 a2 ),
(a1 a2 + a4 )
′′
a3 = a3 . (2.10)

This is not just a coincident but a strong condition which is generally true for
any kind of initial seeds and it says that the process of mutating a quiver twice is an
8 Chapter 2. Cluster Algebras

rule 1

rule 2

rule 3

F IGURE 2.5: Example of a graph.

involution. It can be stated as:

µ2k = Identity (2.11)

With the following basics and definitions at our disposal we are now ready to
construct from the scratch the cluster algebras on our own.

2.3 Cluster Algebras


Lets start form the simplest yet complete example which will make use of all the
definitions discussed yet and will develop thorough understanding of how to imple-
ment them to construct specific algebra. Construction of cluster algebras depends
highly on the initial seeds. Even the dependence is to this extent that the initial seed
is enough to determine if the algebra is going to be finite or infinite. The First algebra
we will develop is the A2 cluster algebra.

2.3.1 A2 Cluster Algebra


The A2 cluster algebra starts from a seed with two nodes connected by a single arrow.
Take a seed say S1 = (a1 , Q1 ) with

a1 = (a1 , a2 ),

 
0 1
Q1 = (2.12)
−1 0

with the following quiver:

S1
a1 a2
2.3. Cluster Algebras 9

F IGURE 2.6: Example of a graph.

We can mutate this quiver in two places. Starting from S1 and mutating it on a1
we get a new seed as

S2 = µ1 (S1 ) ,
S2 = µ1 (a1 , Q1 )
S2 = (a2 , Q2 ) . (2.13)

which can be shown in its expanded cluster, quiver notation as

   0 −1 

S2 = a1 , a2 , (2.14)
1 0

Now we only have a single mutation µ2 which can produce a unique quiver be-
cause mutating again on first node will take us back to the original seed. By applying
mutation µ2 we get the following seed

   0 1 
′ ′
S3 = a1 , a2 , (2.15)
−1 0

The quiver mutation up till now can be seen in Figure 2.6. Lets now first find the
′ ′ ′
two new cluster variables a1 and a2 by using the relation 2.9. The variable a1 was
generated on mutating S1 by µ1 so

1
a′1 = (1 + a2 ) ,
a1
1 + a2
a′1 = . (2.16)
a1

In the same way we can find a2 as

1
a′2 = (1 + a′1 ) ,
a2
1 + a1 + a2
a′2 = . (2.17)
a1 a2
For now it seems that this process of mutation is making our cluster variables
more and more complex, but if one dares to go on further this algebra starts to sur-
render (well at least in some cases), and reveal its simplicity. To proceed we mutate
10 Chapter 2. Cluster Algebras

F IGURE 2.7: Example of a graph.

the first vertex as it is the only mutation which will give us a unique seed. We get S4
′′
and a new variable a1 on first vertex as

   0 −1 
′′ ′
S4 = a1 , a2 , (2.18)
1 0
′′
Now mutating on second vertex we get S5 and another variable a2 at this vertex
given by Equation 2.19, and this sequence of mutations is shown in Figure 2.7

   0 1 
′′ ′′
S5 = a1 , a2 , (2.19)
−1 0

Now lets find the cluster variables of S5 . The operation µ1 (S3 ) produced S4 and
by using equation 2.9 again we get

′′ 1
a1 = (1 + a′2 ) ,
a′1
 
′′ a1 1 + a1 + a2
a1 = 1+ ,
1 + a2 a1 a2
′′ a1 (a1 a1 + 1 + a1 + a2 )
a1 = ,
1 + a2 a1 a2
′′ 1 (1 + a1 )(1 + a2 )
a1 = ,
1 + a2 a2
′′ 1 + a1
a1 = . (2.20)
a2
′′
With the same procedure a2 becomes

′′1  ′′ 
a2 = a + 1 ,
a′2 1
′′ a1 a2 1 + a1 + a2
a2 = ,
1 + a1 + a2 a2
′′
a2 = a1 . (2.21)

Quite remarkably we have recovered one of our cluster variables just by repeated
mutations. Only if we can recover the other variable along with the recovered one
then periodicity will be achieved in true sense. We produced S5 by applying µ2 on
′′
S4 . Lets move further and apply µ1 on S5 . It will mutate the a1 node and produce a
′′′
new variable a1 as
2.3. Cluster Algebras 11

(a)

=
(b)

= =

F IGURE 2.8: Example of a graph.

′′′ 1  ′′

a1 = ′′ 1 + a2 ,
a1
′′′ a2 1 + a1
a1 = ,
1 + a1 1
′′′
a1 = a2 . (2.22)

This concludes that the S6 is actually equal to S1 and in case of the seed we took
initially the algebra was finite. The final mutation and equivalence of S6 and S1 can
be seen in Figure 2.8 (a) and 2.8 (b) respectively.
Final thing to discuss here is the exchange graph which is a beautiful and power-
ful method of expressing the whole structure of an algebra. Whatever we have done
in this example of A2 cluster algebra can be summarized with an undirected graph
which is called an exchange graph. In this example we have total of five distinct
seeds which are linked by each other through a single mutation as

S2 = µ1 (S1 ) ,
S3 = µ2 (S2 ) ,
S4 = µ1 (S3 ) ,
S5 = µ2 (S4 ) ,
S1 = µ1 (S5 ) (2.23)

This forms a closed cycle of mutations and which means that starting from any
seed and any mutation you will get back to the seed you started with. The exchange
graph for A2 is given in Figure 2.9.
12 Chapter 2. Cluster Algebras

F IGURE 2.9: Exchange graph of A2 algebra.


13

Chapter 3

Spinors

One cannot think of starting supersymmetry without knowing about the spinors.
Apart from the theory itself being difficult, the first hurdle one faces in learning
supersymmtery is developing the understanding of Weyl and Majorana spinors and
the intimidating notation in which they are described. Without that, it is really daunt-
ing if not impossible, to learn supersymmtery. This chapter is dedicated on spinors
starting from the usual Dirac spinors and then advancing to the required spinors.

3.1 Dirac Equation and the Dirac Spinor


Dirac equation was an attempt to formulate a relativistic version of the famous Shrodinger’s
Equation introduced by Paul Dirac in 1928. This Equation reads as follows

(ι̇γ µ ∂µ − m)Ψ = 0 (3.1)

where ∂µ is a differential operator and Ψ is the Dirac spinor which is a four compo-
nent spinor. γ µ are the 4 × 4 Dirac matrices which gave the Dirac Equation its true
glory. These matrices can be represented in different ways and the representation we
will follow is
   
0 0 1 0 0 0 0 −1
 0 0 0 1   0 0 −1 0 
γ0 = 1 0 0 0 , γ =  0 1 0
 1  ,
0 
0 1 0 0 1 0 0 0
    (3.2)
0 0 0 ι̇ 0 0 −1 0
 0 0 −ι̇ 0 
 , γ3 =  0 0 0 1 

γ2 =   0 −ι̇ 0 0   1 0
.
0 0 
ι̇ 0 0 0 0 −1 0 0

Using the Euler-Lagrange equations, one can derive the expression for the Dirac
Lagrangian which takes the following form

LDirac = Ψ̄ (γ µ Pµ − m) Ψ. (3.3)

Here Pµ is the four momentum operator which equals ι̇∂µ . The barred spinor rep-
resents the Dirac adjoint which is related to its unbarred spinor via one of the dirac
matrices as
Ψ = Ψ† γ 0 . (3.4)
14 Chapter 3. Spinors

Lastly just for the reference we define another matrix γ 5 as


 
1 0 0 0
 0 1 0 0 
γ5 = 
 0 0 −1 0 
 (3.5)
0 0 0 −1

and it is related to chirality of spinors which we will see in the next section. The
representation we used is called Weyl or chiral representation and the γ 5 matrix is
called chirality operator. It can be represented in terms of the Dirac matrices as

γ 5 = ι̇γ 0 γ 1 γ 2 γ 3 . (3.6)

3.2 Weyl Spinors


Contrary to Dirac spinors, Weyl spinors are two component spinors and are actually
contained inside the Dirac spinor. Starting with the four component Dirac spinor Ψ,
 
ψ1
 ψ2 
Ψ=  ψ3  ,
 (3.7)
ψ4

we define new two component spinors η and χ as


   
ψ1 ψ3
η= χ= . (3.8)
ψ2 ψ4

These spinors are called Weyl Spinors. Substituting this in Equation 3.7, we get
Dirac spinor in terms of Weyl spinors:
 
η
Ψ= . (3.9)
χ

But what is the motivation behind this decomposition of Dirac spinors? Well we
will see in the upcoming sections that under Lorentz transformation of ψ, η and χ
tranforms independently of each other. In general upon any tranformation a vector
components gets mixed with each other. But in case of Dirac spinors, either of the
upper two components which are in η never mixes with any of the lower two compo-
nents in χ. It means that Dirac spinors are reducible while the Weyl spinors are not
reducible and are mathematically more fundamental than Dirac spinors.

As promised, lets get back to chirality and look at the eigenstates of γ 5 . We can
easily create two eigenstates of γ 5 from the Dirac spinor by setting one of the Weyl
3.2. Weyl Spinors 15

spinors equal to zero for each eigenstate. We get the following eigenvalue equations
   
5 η η
γ =+ ,
0 0
   
5 0 0
γ =− . (3.10)
χ χ

These eigenvalues describes the chirality of a spinor. Eigenvalues +1 and −1


corresponds to right-chiral and left-chiral Weyl spinors respectively. Thus η and χ
in our case are right and left chiral spinors respectively. To define it more formally,
we can introduce two projection operators PR and PL which simply project out the
right-chiral and left-chiral Weyl spinors when applied on Dirac spinors as
   
ηR ηR
PR Ψ = PR = ,
χL 0
   
ηR 0
P L Ψ = PL = . (3.11)
χL χL

Applying the chirality operator reveals the chirality of spinors as

γ5 (PR Ψ) = PR Ψ,
γ5 (PL Ψ) = −PL Ψ. (3.12)

Representation of these projection operators is simply


   
1 0 0 0
PR = PL = (3.13)
0 0 0 1

and these operators can be written in terms of chirality operator as


1 + γ5 1 − γ5
PR = PL = . (3.14)
2 2
The question that can be asked now is that if Weyl spinors are more basic than
Dirac spinors, then is there a way of expressing Dirac equation in terms of Weyl
spinors? First hurdle in doing so is clearly visible i.e. that Dirac matrices are 4 × 4
but Weyl spinors are of two components thus the dimensions don’t match. We need
to find some 2 × 2 replacement of the Dirac matrices to accomplish this.

This is done by representing the Dirac matrices in terms of 2 × 2 Pauli matrices.


In this way Equation 3.2 becomes
   
0 0 I 1 0 −σ 1
γ = , γ = ,
I 0 σ1 0
    (3.15)
2 0 −σ 2 3 0 −σ 3
γ = , γ = ,
σ2 0 σ3 0
16 Chapter 3. Spinors

whereas Pauli matrices along with σ 0 are given by


   
0 1 0 1 0 1
σ = , σ = ,
0 1 1 0
   
2 0 −i 3 1 0
σ = , σ = (3.16)
i 0 0 −1

or simply
σ µ = σ 0 , ⃗σ .

(3.17)
In the same way,
γ µ = γ 0 , ⃗γ .

(3.18)
Using the minkowski metric with mostly plus convention, its covariant form becomes

γµ = ηµν γ µ = γ 0 , −⃗γ .

(3.19)

Now finally inserting Equation 3.9 and the form of Dirac matrices given in Equation
3.19 in the Dirac equation we get

(E1 − ⃗σ · p⃗)ηR = mχL


(3.20)
(E1 + ⃗σ · p⃗)χL = mηR .

Look closely and notice that these equations are coupled in left and right chiral
Weyl spinors via a mass m. With a little bit of new notation i.e.

σ µ = (1, ⃗σ ), σ̄ µ = (1, −⃗σ ) (3.21)

we get,  
µ 0 σ̄ µ
γ = , (3.22)
σµ 0
and finally Equation 3.20 becomes

(σ̄ µ · pµ )ηR = mχL


(3.23)
(σ µ · pµ )χL = mηR .

In the end we can describe the Dirac Lagrangian as

LDirac = ηR† σ µ i∂µ ηR + χ†L σ̄ µ i∂µ χL − mηR† χL − mχ†L ηR . (3.24)

3.3 Charge Conjugation


As mentioned in the previous section, supersymmetric theories are written using the
Weyl spinors and Majorana spinors. We have discussed the former but to go further
towards Majorana spinors, it is crucial to first introduce an operation which is applied
on spinors. The operation of charge conjugation relates a field of a particle to its anti
particle field. Lets take the Dirac spinor in Weyl representation. For now and for
the rest of this thesis, we take η to represent the right chiral Weyl spinors and χ to
represent the left chiral Weyl spinors. The Dirac spinor for a certain partical p, which
3.3. Charge Conjugation 17

can be any spin half particle say an electron, quark or a neutrino, can be written as:
 
ηp
Ψp = . (3.25)
χp

Using the charge conjugation operation, we can define its corresponding antipar-
ticle p̄ spinor as
 
c ηp̄
Ψp = Ψp̄ = . (3.26)
χp̄

Here c denotes the operation of charge conjugation. Both ηp and ηp̄ are right chiral
forming a particle-antiparticle pair. Similarly, χp and χp̄ are two different left chiral
particles. We can explicitly define this operation as

Ψcp = C Ψ̄Tp̄ . (3.27)

C is the charge cojugation operator which is taken as

C = −ι̇γ 2 γ 0
  
0 1 0 σ2
C = −ι̇
1 0 −σ 2 0
 2 
ι̇σ 0
C= . (3.28)
0 −ι̇σ 2

Now in Equation 3.27 we write the Dirac conjugate explicitly as


T
Ψcp = C Ψ̄Tp = C Ψ†p γ 0
T
= C γ 0 Ψ†T p = Cγ Ψp
0 †T

Ψcp = C0 Ψ†T
p , (3.29)

where C0 is simplified as
 
0 2 0 0 2 0 iσ 2
C0 = Cγ = −iγ γ γ = −iγ = . (3.30)
−iσ 2 0

Inserting this in Equation 3.29 we get


     †T 
c 0 iσ 2 †T 0 iσ 2 ηp
Ψp = Ψp = . (3.31)
−iσ 2
0 −iσ 2
0 χ†T
p

Comparing it with Equation 3.25 we get the following relations

ηp̄ = iσ 2 χ†T
p
(3.32)
χp̄ = −iσ 2 ηp†T .

We have successfully shown here the relation between the particle spinors and
their antipartle spinors. We know that ηp and χp are right and left chiral spinors.
18 Chapter 3. Spinors

But what about their antiparticles? One might think that since ηp̄ is made of a left
chiral spinor as shown in Equation 3.32, therefore ηp̄ itself will also be a left chiral
spinor. However, that is not the case. Surprisingly the antiparticle spinor ηp̄ is right
chiral spinor and χp̄ is left chiral spinor although they are made of right chiral particle
spinor and left chiral particle spinor respectively. You can see its justification in the
next chapter which uses the Lorentz transformation of spinor to find out the chirality
of antiparticle spinors.

In the same way, by applying charge conjugation to Equation 3.29 we get

ηp = iσ 2 χ†T

(3.33)
χp = −iσ 2 ηp̄†T .

3.4 Majorana Spinors


Now we are ready to discuss the Majorana spinors which are four component spinors
like Dirac spinors, but with the difference that they remain unchanged by charge
conjugation.
   
c ηp ηp̄
ΨM = ΨM̄ = = . (3.34)
χp χp̄

We have identified ηp = ηp̄ and also χp = χp̄ which is possible only if the
particle is its own antiparticle. This type of field does not contain any quantum
numbers which can differentiate the particle from its antiparticle like a lepton number
or electric charge. Interestingly, with only one of the spinors, either left chiral or
right chiral ones, identified with its charge conjugated partner, we can prove that the
rest are identical to each other. As an example just take ηp = ηp̄ then by charge
conjugation formula in Equation 3.32, we get ηp = ι̇σ 2 χ†Tp . Now by inserting this in
Equation 3.34 we get
   2 †T 
ηp iσ χp
ΨM = = . (3.35)
χp χp

We have clearly shown here that antiparticle left chiral state is actually the particle
left chiral state by equating the right chiral states of particle and antiparticle with each
other. By inserting Equation 3.35 in Dirac Lagrangian, we get the Lagrangian for free
Majorana particle as
1
LM = Ψ̄M (γ µ i∂µ − m) ΨM . (3.36)
2
19

Chapter 4

Lorentz Invariants and


Transformations

In our current understanding of physics, we know that every theory must be Lorentz
invariant. In simple words, it ensures that the physics explained by the theory is in-
dependent of the inertial frame that we choose, regardless of their relative speeds and
orientation. To see how we can make Lorentz invariants using our spinors, we will
discuss how spinors are transformed under these transformations. We will introduce
a notation which is very useful in supersymmetry and then develop some Lorrentz
invariants out of Weyl spinors. This notation may seems a bit annoying at first but
supersymmetry gets very complicated very quickly and the notation helps us survive
in the intricacy of the theory.

4.1 Lorentz Transformation of Spinors


Spinors don’t transform like scalars, vectors or tensors. It is evident by the Dirac la-
grangian that spinors must transform in a specific way under Lorentz transformations
because ∂µ in its kinetic term is a covariant vector and the γ µ are fixed (which is the
chiral representation we used in the previous section). Thus, Ψ must transform in a
way that it makes the whole term invariant. Instead if we consider Ψ to be scalar then
γ µ must not be fixed which means that they will not be the same for stationery and
moving observers. This can not be the case and thus we can clearly see that Ψ is not a
scalar or a vector and still it transforms appropriately under Lorentz transformations.

Assume that spinors transform under Lorentz transformations as


′ ′
Ψ (x ) = SΨ(x) (4.1)

the matrix s is a 4 × 4 matrix and we require that it must be invertible so its inverse
transformation must also exist, which gives us
′ ′
Ψ(x) = S −1 Ψ (x ). (4.2)

The matrix which gives the proper transformation of Ψ is


" #
i 1⃗
⃗ϵ · ⃗σ − 2 β · ⃗σ 0
S = exp 2 i 1⃗ , (4.3)
0 2

ϵ · ⃗
σ + 2
β · ⃗
σ
20 Chapter 4. Lorentz Invariants and Transformations

where ⃗ϵ is the infinitesimal rotation vector and β is the infinitesimal boost parameter.
The direction and magnitude of ⃗ϵ gives us the direction of axis of rotation and the
amount of rotation respectively. The derivation of this transformation matrix is a bit
involved and if you’re interested you can see appendix***. On expanding it to first
order we get
 
′ i 1⃗
ηR → ηR = 1 + ⃗ϵ · ⃗σ − β · ⃗σ ηR ,
2 2
  (4.4)
′ i 1⃗
χL → χL = 1 + ⃗ϵ · ⃗σ + β · ⃗σ χL .
2 2

We can clearly see here that left and right chiral spinors transforms independently
which means that a Dirac spinor is actually reducible and therefore is described by
the irreducible representation of Weyl spinors.

We can now start from the Dirac Lagrangian and look for Lorentz invariant quan-
tities. Take the mass term for now which says that Ψ† γ 0 Ψ is a Lorentz invariant
quantity. Writing it in terms of Weyl spinors we get

Ψ† γ 0 Ψ = η † χ + χ† η. (4.5)

Thus we have a Lorentz invariant in terms of Weyl spinors and it turns out that
both terms on the right are individually invariant which you can easily check by
applying transformation shown in Equation 4.2.

4.2 Lorentz Invariants


We have just developed two invariants using the Dirac Lagrangian. Now we wish to
develop more invariants out of Weyl spinors. We have seen how η and χ transform
under Lorentz transformation but one can ask what is the transformation of η † and
χ† ? It must be clarified here that the hermition conjugation applied to the spinors
here is both an operation of linear algebra and an operation of quantum field theory.
It means that not only the spinor components are hermition conjugated but also the
spinor is transposed in linear algebra sense. So a spinor, say χ† , represents a row
vector with its components being χ†i given as

χ† = χ†1 χ†2 .

(4.6)

So to make it a column vector we will take its transpose


 † 
†T χ1
χ = . (4.7)
χ†2
4.2. Lorentz Invariants 21

Lets look at the transformation of this object by applying Equation 4.4, we get
 ′
†T
χ†T
L → χ†T
L = (χ′L )
 ∗
i 1⃗
= 1 + ⃗ϵ · ⃗σ + β · ⃗σ χ†T L
2 2
 
i 1
= 1 − ⃗ϵ · ⃗σ ∗ + β⃗ · ⃗σ ∗ χ†TL (4.8)
2 2

It can be seen here that this doesn’t transform like a spinor. But quiet surprisingly
if we add the factor of ι̇σ 2 and apply the transformation on the new object, it then
follows one of the transformations in Eqaution 4.4:
′  ∗
2 †T

2 †T 2 ∗ i 1⃗
1 + ⃗ϵ · ⃗σ + β · ⃗σ χ†T

iσ χL → iσ χL = iσ L
2 2
 
i 1
= iσ 2 1 − ⃗ϵ · ⃗σ ∗ + β⃗ · ⃗σ ∗ χ†T
L
2 2
 
i 1
= 1 + ⃗ϵ · ⃗σ − β⃗ · ⃗σ iσ 2 χ†T L , (4.9)
2 2

which is the transformation of a right chiral spinor. We have discovered here an


object which transforms as a right chiral spinor but is composed of a left chiral spinor.
In the same way you can check that we can build an object out of right chiral spinor
which transforms like a left chiral spinor. Both of these objects are given as

ι̇σ 2 χ†T transforms as right chiral spinor


−ι̇σ 2 η †T transforms as left chiral spinor (4.10)

Now lets get back to our discussion of builiding Lorentz invariant quantities. We
know from the previous section that χ† η is a Lorentz invariant. We also know that
both η and ι̇σ 2 χ†T transforms as right chiral spinors so it means that we can replace
former by later in the invariant χ† η and the new quantity formed is still invariant:

(ι̇σ 2 χ†T )† χ = χT (ι̇σ 2 )† χ = χT (−ι̇σ 2 )χ. (4.11)

In the same way by replacing χ with −ι̇σ 2 η †T in η † χ which is an invariant, we get


another invariant

η † (−ι̇σ 2 η †T ) = −η † ι̇σ 2 η †T . (4.12)

This shows that we can express a Lorentz invariant in terms of either of the two
chiralities of the spinors. We now introduce two types of dot products which serve as
a shorthand notation and will be used in the later sections. This dot product is used to
describe the Lorentz invariants. We define the dot product of two left chiral spinors
as

χ · χ ≡ χT (−ι̇σ 2 )χ (4.13)
22 Chapter 4. Lorentz Invariants and Transformations

and the other one as

χ̄ · χ̄ ≡ χ† ι̇σ 2 χ†T . (4.14)

4.3 Supersymmetric Transformations of Fields


Supersymmetry is a global symmetry between two differnet types of particles. It
is a spacetime symmetry between Bosons and Fermions which have integral spins
and half integral spins respectively. The equations of matter and forces are identical
in supersymmetric theories. This is the idea of supersymmetry that the force fields
(Bosonic fields) are transformed into matter fields (Fermionic Fields) and vice versa.
So we take scalar field and say that the variation of scalar field is proportional to the
fermionic field as

ϕ = ϕ + δϕ (4.15)

and the variation of field δϕ is proportional to χ. The expression for variation of


scalar field turns out to be

δϕ = ζ · χ, (4.16)

where ζ is an infinitesimal constant Weyl spinor which takes different values in dif-
ferent inertial frames
 
ζ1
ζ= . (4.17)
ζ2

In the same way variation of χ includes scalar field ϕ and its expression is

δχ = −ι̇(∂µ ϕ)σ µ ι̇σ 2 ζ ∗ . (4.18)

In this way the transformations of variation of hermition conjugate of fields are given
by

δϕ† = ζ̄ · χ̄ (4.19)

and

δχ† = −ι̇(∂µ ϕ† )ζ T ι̇σ 2 σ µ . (4.20)

The the parameter ζ we have taken is infinitesimal and a constant parameter for
a particular frame. This is called global supersymmetry. But if we consider ζ to
be spacetime dependent then it forces us to introduce a gauge field. Under all these
supersymmetric transformations the Lagrangian remains invariant.
23

Chapter 5

Generators of Supersymmtery

Whenever there is a conserved quantity there is a symmetry. Symmetry of a system


tells us that the system is invariant under some transformations and these transfor-
mations are carried out by Generators. For example conservation of momentum in a
system holds for every inertial observer out there. It means that the system is invari-
ant under translations which are generated by momentum generators. Likewise if a
system is invariant under time translations then energy of that system is conserved
and these are carried out by energy generators. These generators are often called
charges.

5.1 Generators and Their Algebra


A system is defined by a Lagrangian. If a Lagrangian is symmetric under some trans-
formations then it is natural to ask what type of charges are producing that symmetry.
In this chapter we will find out what type of charges produce supersymmetry and also
that the algebra that the charges obey is more useful than the charges themselves.
Usually this algebra is defined by the commutation and anticommutation relations of
charges and is called Lie algebra and graded Lie algebra respectively.

5.1.1 General Discussion


Consider the canonical quantization and think of the scalar field operator ϕ. To eval-
uate any physical process we need to find the expectation values of the field

⟨a| ϕ |b⟩ . (5.1)

If our theory containing ϕ has a symmetry, then the expectation values must remain
unchanged under transformations related to that symmetry. Taking our transforma-
tion to be a spacetime transformation then
′ ′ ′
⟨a | ϕ(x ) |b ⟩ = ⟨a| ϕ(x) |b⟩ (5.2)

and the transformed states are related to the previous ones as


′ ′
|b ⟩ = U † |b⟩ ⟨a | = ⟨a| U. (5.3)
24 Chapter 5. Generators of Supersymmtery

Upon inserting this in Equation 5.2 we get



⟨a| U ϕ(x )U † |b⟩ = ⟨a| ϕ(x) |b⟩ ,
′ ′
⟨a| ϕ (x ) |b⟩ = ⟨a| ϕ(x) |b⟩ . (5.4)
′ ′
This says that ϕ (x ) = ϕ(x) and we can conclude this result as

ϕ (x) = ϕ(x + δx). (5.5)
′ ′ ′
Also in Equation 5.4 we have made a substitution ϕ (x ) = U ϕ(x )U † which we
restate as

ϕ (x) = U ϕ(x)U † . (5.6)

Here comes the role of generators which generate the transformations. We as-
sume that our transformations are parameterized by n infinitesimal parameters which
are ϵ1 , ϵ2 , ϵ3 , ...ϵn . Then there must be the same number of generators Q1 , Q2 , Q3 , ...Qn .
Transformation U then in terms of charges is given by

U = e±ι̇ϵ·Q , (5.7)

where this dot product is the short notation for summation ϵi Qi . The parameter ϵ
here could be anything like constant numbers or Grassmann numbers such as con-
stant Weyl spinors which will be used in later sections as an infinitesimal parameter
of supersymmetric transformations. So based on transformations and their parame-
ters, we will have our bosonic and fermionic generators.

Now by substituting Equation 5.7 in Equation 5.6 we get



ϕ (x) = e±ι̇ϵ·Q ϕ(x)e∓ι̇ϵ·Q . (5.8)

Expanding it to the first order of ϵ we get

ϕ′ (x) ≈ (1 ± iϵ · Q)ϕ(x)(1 ∓ iϵ · Q) (5.9)


= ϕ(x) ± iϵ · Qϕ(x) ∓ iϕ(x)ϵ · Q (5.10)
= ϕ(x) ± i[ϵ · Q, ϕ(x)]. (5.11)

The total change in the field δϕ induced after transformation can be written as

ϕ (x) − ϕ(x) = δϕ(x). (5.12)

Comparing it to Equation 5.9, we arrive at a cucial result

±[ϵ · Q, ϕ] = −ι̇δϕ(x). (5.13)

This is a powerful result which relates charges of transformations of fields with


the coordinate transformation in a simple commutator expression.
5.1. Generators and Their Algebra 25

5.1.2 Generators and Algebra of Supersymmetry


In the previous chapter we discussed the SUSY transformations in which we had a
constant Weyl spinor and also its conjugate as infinitesimal parameters. This directly
indicates that there must be four generators which we say are Q1 , Q2 , Q†1 and Q†2 .
These generators are themselves components of Weyl spinors and we categorize them
as
   † 
Q1 † Q1
Q= Q = . (5.14)
Q2 Q†2

Now that we have the required charges we can formulate the unitary operator U
which transforms the fields. We now know that the argument of the exponential in
operator U must be a Lorentz invariant so we have to combine the charges and the
parameters in such a way that they fulfil the condition. As a convention we take Q and
ζ to be left chiral spinors and from the previous chapter, we know the combination
to build invariants out of them. Using Equations 4.13 and 4.14 we get two Lorentz
invariants

Q · ζ = Q(−ι̇σ 2 )ζ Q̄ · ζ̄ = Q† ι̇σ 2 ζ ∗ . (5.15)

Now we can write the unitary operator for SUSY transformations in terms of
charges as

Uζ = eι̇Q·ζ+ι̇Q̄·ζ̄ . (5.16)

By applying Equation 5.13 to the supersymetric transformations discussed in the


section 4.3 of previous chapter, we get the following commutation relations

[ζ · Q, ϕ] = −ι̇ζ · χ,
 
ζ̄ · Q̄, ϕ] = 0,
[ζ · Q, χ]] = 0,
ζ̄ · Q̄, χ] = −ι̇(∂µ ϕ)σ µ ι̇σ 2 ζ ∗ .
 
(5.17)

Considering another transformation Uβ = eι̇Q·β+ι̇Q̄·β̄ and applying Uβ and Uζ on


ϕ we get

Uβ Uζ ϕUζ Uβ = δβ δζ ϕ. (5.18)

The commutator of these two transformations become

Uβ Uζ ϕUζ Uβ − Uβ Uζ ϕUβ Uζ = δβ δζ ϕ − δζ δβ ϕ. (5.19)

Expanding the exponential form of U to the first order we get


  
δβ δζ ϕ − δζ δβ ϕ = − Q · β + Q̄ · β̄, Q · ζ + Q̄ · ζ̄, ϕ
  
+ Q · ζ + Q̄ · ζ̄, Q · β + Q̄ · β̄, ϕ . (5.20)
26 Chapter 5. Generators of Supersymmtery

Now using the Jacobi identity we can write it as


  
δβ δζ ϕ − δζ δβ ϕ = Q · ζ + Q̄ · ζ̄, Q · β + Q̄ · β̄ , ϕ (5.21)

and solving further gives us four commutators as


  
δβ δζ ϕ − δζ δβ ϕ = [[Q · ζ, Q · β] , ϕ] + Q · ζ, Q̄ · β̄ , ϕ
     
+ Q̄ · ζ̄, Q · β , ϕ + Q̄ · ζ̄, Q̄ · β̄ , ϕ . (5.22)

To go further we expand the spinors in the commutators in their components forms


and we get the following
ab cd
[Q · ζ, Q · β] = σ 2 σ2 ζb βd {Qa , Qc }
2 ab cd
σ 2 ζb βd∗ Qa , Q†c
  
[Q · ζ, Q̄ · β̄] = − σ
ab 2 cd ∗  †
[Q̄ · ζ̄, Q · β] = − σ 2 σ ζb βd Qa , Qc
ab cd
[Q̄ · ζ̄, Q̄ · β̄] = σ 2 σ 2 ζb∗ βd∗ Q†a , Q†c .
  
(5.23)

A noticeable thing here is that we started with the commutation relations but
ended up with anticommutation relations (graded Lie algebra). This is due to the
Grassmann nature of infinitesimal parameters. We haven’t done anything to the left
hand side of 5.22. Using the SUSY transformation from section 4.3 we get

δβ δζ ϕ = δβ −ζ T iσ 2 χ


= −ζ T iσ 2 δβ χ
= ζ T iσ 2 iσ µ iσ 2 β ∗ ∂µ ϕ
= −iζ T σ 2 σ µ σ 2 β ∗ ∂µ ϕ
= −iζ T σ̄ µT β ∗ ∂µ ϕ
δβ δζ ϕ = iβ † σ̄ µ ζ∂µ ϕ. (5.24)

Likewise, applying transformations on δζ δbeta and inserting them in left hand side of
Equation 5.22 we get

δβ δζ ϕ − δζ δβ ϕ = −i ζ T σ 2 σ µ σ 2 β ∗ − β T σ 2 σ µ σ 2 ζ ∗ ∂µ ϕ.

(5.25)

Substituting the ∂µ ϕ = [Pµ , ϕ] in the above expression we get

δβ δζ ϕ − δζ δβ ϕ = ζ T σ 2 σ µ σ 2 β ∗ − β T σ 2 σ µ σ 2 ζ ∗ [Pµ , ϕ]


= ζ T σ 2 σ µ σ 2 β ∗ − β T σ 2 σ µ σ 2 ζ ∗ Pµ , ϕ .
  
(5.26)

Now by comparing Equations 5.22 and 5.26 we get the final anticommutation rela-
tions

{Qa , Qb } = 0 (5.27)

{Q†a , Q†b } = 0 (5.28)


5.2. Nonclosure SUSY Algebra and the Need of an Auxilary Field 27

{Q†a , Qb } = 2(σ µ )ba Pµ (5.29)

{Qa , Q†b } = 2(σ µ )ab Pµ (5.30)

We have shown here the anticommutation relations of the SUSY charges. What
we are missing to complete the SUSY algebra is the commutation relations of these
charges with the Poincare generators which are Pµ , the generator of translations and
Mµν , the generator of Lorentz transformations. We know by the operator form of Pµ
that is acts on the functions of positions. By looking at the SUSY transformations on
ϕ, we know that SUSY charges don’t produce any functions of position which means
that SUSY generators commutes with Pµ .

[Qa , Pµ ] = 0 (5.31)

[Q†a , Pµ ] = 0 (5.32)

The working of commutation relations with Lorentz generators is not this sim-
ple. The detailed derivation can be seen in (appendix/reference)*** and the relation
becomes

[Q̄ȧ , Mµν ] = −Q̄ḃ (σ̄µν )ḃȧ (5.33)

[Q̄ȧ , Mµν ] = (σ̄µν )ȧḃ Q̄ḃ (5.34)

5.2 Nonclosure SUSY Algebra and the Need of an Aux-


ilary Field
The algebra we have developed so far is by using the scalar field ϕ. To generalize
the treatment, we have to prove that the same algebra is also followed if we use the
spinor field χ. But this is not the case with the SUSY transformations we have chosen
in section 4.3. This algebra is just valid for the scalar field and is not closed when
we use the spinor field. To make it closed for both fields, we introduce an Auxiliary
Field F which modifies the spinor variation in such a way that the same algebra is
also valid for it. In presence of this field the SUSY transformations become

δϕ = ζ · χ, (5.35)

δχ = −ι̇σ µ (ι̇σ 2 ζ ∗ )∂µ ϕ + F ζ, (5.36)

δF = −ι̇ζ ∗ σ̄ µ ∂µ χ. (5.37)

The nonclosure of the SUSY algebra without an auxiliary field can also be viewed
in this way that supersymmetry requires that their must be equal number of fermionic
28 Chapter 5. Generators of Supersymmtery

and bosonic degrees of freedom. We know that a complex scalar field has two degrees
of freedom. Also a spinor field constrained to its equation of motion (on shell spinor)
also has two degrees of freedom thus algebra closes on shell. On the other hand a
Weyl spinor field without any restriction to follow any equation of motion (off shell
spinor) has two complex components and four real components thus four degrees of
freedom which do not match with scaler field. It means that algebra does not close
off the shell. By introducing an auxiliary field which has zero on shell degrees of
freedom but two off shell degrees of freedom we can balance the number and get
closed algebra in both off shell and on shell cases.

You might also like