Universit Catholique de Louvain Facult des Sciences LPHY 2125 - Relativistic Quantum Mechanic Fabio Maltoni

First Steps to the Supersymmetry

Louvain - La - Neuve, Anne acadmique 2011-2012

Notes
This work is a summary of the rst three chapters of "Supersymmetry and the MSSM : An Elemantary Introduction". Notes of lectures for graduate students in Particle Physics (Oxford 2004 & 2005) written by Ian J. R. Aitchison. I chose to write this resume in English because I think it is a good exercise. Unfortunately, I do not master English very well. Therefore, I apologize in advance for the mistakes this report may(will) contain.

Introduction and Motivation of Supersymmetry

First of all, we can notice that a dominant theme in twentieth century physics was that of symmetries. We sought what was the symmetries respected by the dierent interactions. That work has taught us so much about them that today the interactions are characterized by the symmetries they verify. We checked principally the parity (P), the time reversal (T) and the charge conjugation (C). It is natural as a physicist to ask if our current quantum eld theories of the Standard Model (SM) exploit all the kinds of symmetry which could exist, consistent with Lorentz invariance. At which one could reply, "Why would you spend your time seeking to modify the SM while it is not broken by the experiment ?". Indeed, it works even very well ! Nevertheless, history tell us that exploring the imperfections of a theory can lead to great discoveries. And the SM has some imperfections : it doesnt impose any constraint on the Higgs mass, the gauge coupling doesnt meet convincingly at high energy, there is the "strong CP problem", ... Another imperfection is a "hierarchy problem" : in the electroweak theory, the Higgs eld is responsible for the mass of the gauge bosons and fermions by a mechanism of spontaneous symmetry breaking (BEH Mechanism). The Higgs potential is (1) V = 2 + ( )2 4 where > 0 is the intensity of the Higgs self-interaction and 2 > 0 is a mass term. When we try to compute the self-interaction of the Higgs at oneloop order using the potential (1) we end up with an integral which diverges

1 INTRODUCTION AND MOTIVATION OF SUPERSYMMETRY

quadratically due to the term. To avoid such a non-sense we have to produce a correction to the 2 term where is called the "cuto". The Lagrangian parameter can be chosen to depend on the cut-o in such a way that, when the nal (renormalized) coecient of has the desired value. In term of physics this trick (the "ne-tuning") is unsatisfactory because it doesnt explain anything. It is similar to an effective potential. With the supersymmetry(SUSY), we must also take into 2 2 account the fermion loop correction gf 2 . So if for some reason = gf , the two-loop corrections cancel each other and we have no longer a quadratic divergence which solves the hierarchy problem. Lets now recall the main ideas of SUSY : for every type of boson, there exists a corresponding type of fermion with the same mass and intrinsic quantum number and vice-versa. It means that every particle in the SM would have a "superpartner" call the sparticle which belong to the other class. Here are a few examples : quark photon gluon (s = 1/2) (s = 1) (s = 1) squark (s = 0) (2)

photino (s = 1/2) gluino (s = 1/2).

To achieve such a result we need a symmetry (the SUSY one) relating bosonic and fermionic degrees of freedom (DOF). We will come back on this point during the next chapter. The algebra of SUSY leads us to say that the space-time coordinates are extended to include further degrees of freedom. The whole new spacetime being called "superspace". A last important thing is that most standard sources on SUSY use either the "dotted and undotted" 2-component spinor notation found in the theory of representations of the Lorentz group, or 4component Majorana spinors. Again, we will discuss spinorial representations in the next chapter.

2 SPINORS

Spinors

There are dierent ways of classifying particles. One of them (the one we are interested in) separates particles in two classes : fermions and bosons. Fermions constitute the matter and bosons carry the interactions. All these particles are associated with a eld which generalises the concept of wave function. The Lagrangian of a system involves a set of elds r and we want it to be invariant under innitesimal transformation such as : r = i rs s where we sum on the s, the rs are some constant coecients and innitesimal parameter. (3) is an

As it is explained above, we want a symmetry which transforms fermionic elds into bosonic ones (and conversely) and which conserves the Lagrangian. It will be of the form : (4) where is an innitesimal parameter, is a fermionic eld and is bosonic eld. A rst consequence of such a transformation is that the number of DOF of bosonic elds have to be the same then the one of fermionic elds. We expected this because we want as many fermions as bosons. The structure of the eld describing a fermion is dierent then the one describing a boson. While we can describe a neutral scalar boson with a one-real-component eld, a fermionic eld is a spinor with at least two components. To avoid such an incoherence we will deal with pseudo-scalar bosons which are charged particles described by a complex (i.e. "two-component") eld. But there is still a problem : we are used to deal with the Dirac spinor and it has four components, not two... Our rst job is therefore to deconstruct the Dirac eld and understand the nature of the two dierent two-component elds wihch constitute it. We will see soon that the dierence has to do with the way they transform under Lorentz transformations. We know that a scalar eld such as the one in the left hand side (LHS) of (4) is invariant under Lorentz transformations. The point is now "how to turn the RHS of (4) Lorentz invariant ?" i.e. "how to put and together properly so as to form a Lorentz invariant ?".

2 SPINORS

2.1

Spinors and Lorentz Transformations

In the following, we use the notation used in standard relativistic quantum mechanics courses dealing with the Dirac equation. Lets do a bit of formalism. Using = c = 1 the Dirac equation is given by : E = ( p + m) with the representation used in small mass or high energy limit :

(5)

0 1 = 1 0

(6)

where i s are the three Pauli matrices. We have also 0 = 0 If we write = the Dirac equation becomes (E p) = m (E + p) = m (9) (10)

0 5 = . 0 1

(7)

(8)

where we see that when m 0, and become eigenstates of the helicity operator p/|p| with eigenvalues +1 and -1 respectively. If m = 0, and are not pure helicity eigenstates but are eigenstates of 5 :
0 5 = 0

5 = 0 0

(11)

The eigenvalue of 5 is called chirality. Eigenstates with chirality +1 are called -type or unfortunately 1 R-type for right-handed spinors. Those with
1. We use the term "right-handed" or "left-handed" also to qualify the helicity and we have just seen that chirality is helicity ONLY in the massless case.

2 SPINORS

chirality -1 are called -type or L-type for left-handed. In SUSY, one works with the two-component and which have well-dened Lorentz transformation character. Lets now consider an innitesimal boost with velocities and rotations with the three innitesimal parameters . Under this Lorentz transformation, and become

= (1 + i /2 /2) = V = (1 + i /2 + /2) = V
1

(12) (13)

where we have dened V = 1 + i /2 /2 which gives V 1 = 1 i /2 + /2 (15) (14)

since the inverse transformation goes in the opposite direction for the boost and for the rotation. The Hermiticity of the s gives : V = 1 i /2 /2. And nally : V
1

(16) (17)

= V 1 = 1 + i /2 + /2.

The equation (12) and (13) tell us (at least) two important things : we know how do and transform under the Poincar group and we see that they behave the same way under rotations but oppositely under boosts. This completes our "rst job" to deconstruct the four-component Dirac spinor and justify the choice to work with this two two-component spinors as a two component one. Now that we know how do and transform under the Poincar group, we can try to construct invariants and 4-vectors.

2 SPINORS

2.2

Constructing Invariants and 4-vectors out of 2component Spinors

It is easy to show that is a Lorentz invariant (i.e. scalar) and that = (, ) = (1, ) transforms as a 4-vector. Using the decomposition of as a bispinor =

(19)

(20)

we see that (18) yields immediately to the conclusion that and are Lorentz invariants. By the same manner we obtain with (19) that ( , ) and ( , ) = transforms as 4-vectors. NB : we introduced the new quantities (1, ), = (1, ). (23) (22) (21)

This last remark shows that it is possible to form 4-vectors with only or only ! Lets now try to form invariants using only or . Looking at the Dirac equation (9) we see that the Lorentz transformation character of p is the same as that of which means that p converts Rtype objects to L-type ones. Looking now at (10) we see that p transforms a L-type object in a R-type one. Considering the complex conjugate of in equation (13) we see also (using the properties of the Pauli matrices) that : 2 = V 2 (24)

which means that 2 transforms like a (see eq. (12)). One can now construct a bispinor with only s and use (18) to form the Lorentz invariant : (i2 (1) ) (2) . (25)

2 SPINORS

We now know how to form the Lorentz invariant scalar product out of two s. This point lead us to rene our notations : the components of -type spinors will have lower indices. From this (lower indices) object we can dene the "upper indices -components" with :

2 . i2 = 1 2

(26)

Multiplying (24) by i and taking its complex conjugate, we see that i2 transforms by V since i2 is a real matrix. We can do exactly the same game with and show that i2 transforms like so that (using again 18) (i2 )T is an invariant. To denote the components of -type spinors we use upper dotted indices a and we dene the lower dotted components with :

1

i2 =

(27)

where a transforms by V 1t .

2 SPINORS

2.3

Majorana Fermions

We can, consistently with the Dirac theory, assemble and i2 into a 4-component object : = M i2 (28)

since i2 transforms like a -type spinor. We have that behaves M under Lorentz transformation like an ordinary Dirac spinor but has fewer degrees of freedom. Indeed, we have only the DOFs of which is a complex two-component spinor. This leads to four DOFs while there is two times more in a Dirac spinor. We can physically understand this particularity by considering the action of charge conjugation operator : c = C0 where C0 = i 2 . We obtain then :

(29)

,c = M

0 i2

i2 0

i2

= = M i2

(30)

which means that describes a spin-1/2 particle even under charge conjuM gation i.e. a particle which is its own antiparticle. For an ordinary fermions, we we have four dierent physical objects : for instance e+ , e , e+ , e . We L L R R see that a Majorana fermion has only the two chirality DOFs. NB : We have also that c i2 ,c = (31) = M M and see easily that a mass term or give us two Lorentz invaM M M M riant. Lets now look at the dot product of two dierent Majorana fermions : 1M 2M = 2M . (32) 1M We know from (29) and (30) that 1M = i 2 which immediately gives 1M 1M 2M = t (i 2 )2M . We conclude therefore that the matrix 1M

i 2 =

i2 0

0 i2

(33)

acts as a metric for the dot product of two Majorana spinors.

2 SPINORS

10

2.4

Dirac Fermions Using L-type Spinors

In a SUSY context, it is very convenient to be able to use only one kind of spinor. By convention, we will use only -type ones (probably because weak interaction acts only on left handed spinors). For instance, instead of using (e ) the right-handed electron eld R for the top component we can use the charge conjugate of the left-handed positron eld :

(e

i2 L

(e+ )

(e ) L

(34)

and we know that its Lorentz transformation character is OK. We will write the top component more compactly : e+c . We see that it is dierent from a L Majorana spinor since its charge conjugate is

(e

)c

i2 L

(e )

(e+ ) L

= . (e ) L

+c eL

(35)

We can therefore conclude that a Majorana fermion is a particle which is its own antiparticle while a Dirac fermion is not its own antiparticle. NB : all the fermions we know today are Dirac ones.

3 A SIMPLE SUPERSYMMETRIC LAGRANGIAN

11

A Simple Supersymmetric Lagrangian

We are now able to construct the simplest supersymmetric theory involving a complex scalar (spin-0) eld and a L-type spinor , both massless. The Lagrangian density for this system is : L = + i . (36)

We know from the Klein-Gordon theory (which concerns spin-0) that the equation respected by is = 0 where = while that for is i = 0 (see (10)). We want to nd a transformations linking and and letting L invariant or unless letting the physics invariant i.e. changing L by adding a total derivative term. Lets discuss the dimensions of what we wrote above. The action is the integral of L over the whole space-time. In the SUN ( = c = 1) a length has the same dimension of a time : M 1 where M is a mass or an energy (which are equivalent since E = mc2 and c = 1). Since the action has to be dimensionless we require that [L] = M 4 . And since the derivatives have the dimension of a mass (or energy) we can deduce that [] = M [] = M 3/2 .

We have already said in (4) that the SUSY transformation has the form : (37)

where we will take independent of the coordinate x. A rst consideration is that the LHS is Lorentz invariant and so has to be the RHS. This is not a problem looking at the work done in previous sections, we just have to declare that is also a -type spinor and form the Lorentz invariant as in the section (2.2) which lead to : = t (i2 ) = 1 2 + 2 1 Looking for the "converse" relation we nd : a = A[i (i2 )]a where A is a constant that we will choose to let the physics invariant. (39) (38)

4 VERY SHORT RESUME

12

Since the Lagrangian involves and we are interested in nding the expression of and : = (1 2 + 2 1 ) = (i2 ) = A t i2 i We are now able to look at L using the Leibniz rule : L = ( ) + ( ) + ( )i + i ( ) T = ( (i2 ) ) + ( (i2 )) (43) T +A( i2 i )i + A i (i i2 ) Putting together the terms involving we see that if we chose A = 1 we get L| = ( i2 ) (44) where we have used = (0 + )(0
2 ) = 0 2

(40) (41) (42)

= .

(45)

Imposing A = 1 to the term involving t allow us to write them as a the sum of two total derivatives so that nally L = i2 + T (i2 ) + T i2 . (46)

We conclude from this last observation that the transformation we "manually construct" is a supersymetric one because it links a fermionic eld and a bosonic eld keeping the action unchanged 2 .

Very Short Resume

We were looking for a transformation which links bosonic eld and fermionic eld. We had to face a big problem : our fermionic elds are Dirac spinors while a bosonic eld is a Lorentz invariant complex eld. Therefore we had to reduce the number of DOFs of the Dirac spinnor and to learn how to create Lorentz invariants with these "lower DOFs spinors". As soon as it was done, we constructed the desired transformation very naturally. And, by adjusting a constant, we saw that it was a symmetry for the action of a system with a complex spin-0 eld and an L-type spinor eld !
2. It is important to say that this last statement is right (for evident reason) only and only if the quantity inside the total derivative vanishes at the boundary of space-time.