An Introduction to Supersymmetry
Frederick McDonnell
Mathematical Sciences Dept., Durham University
Supervisors: Prof. Wojtek Zakrzewski & Prof. Richard Ward
April 27, 2016
Abstract
This report aims to provide an introduction to Supersymmetry (SUSY) for those with limited knowledge of field theory. The Klein-Gordon field is discussed and then quantised, with infra- red and ultra-violet divergences examined. We then discuss the Lorentz group, considering representations of its associated Lie algebra, before extending it to the Poincar´egroup with the inclusion of translations. The Dirac field is then introduced from a historical perspective, and quantised in analogy to the Klein-Gordon field. The concept of Supersymmetry is introduced as an extension of the Poincar´ealgebra, and the relations of the resulting Super-Poincar´ealgebra are examined. We then construct representations of this algebra, called supermultiplets. We discuss the massless free Wess-Zumino model, a supersymmetric theory consisting of a massless chiral supermultiplet (a Weyl fermion and a bosonic complex scalar field). We confirm the action is invariant for this model, and recover the SUSY algebra by inserting quantised fields. Contents
1 Introduction 4 1.1 Preamble ...... 4 1.1.1 Introduction ...... 4 1.1.2 Conventions ...... 5
2 Field Theory 6 2.1 Classical Field Theory ...... 6 2.1.1 Background ...... 6 2.1.2 Equations of Motion ...... 6 2.1.3 Noether’s Theorem ...... 7 2.1.4 An example of Noether’s Theorem ...... 8 2.1.5 On-shell and Off-shell ...... 9 2.2 Moving to Quantum Field Theory ...... 9 2.2.1 The Klein-Gordon Equation ...... 9 2.2.2 Conjugate Momentum and Hamilton’s Equations ...... 10 2.2.3 The Stress-Energy Tensor ...... 10 2.3 Canonical Quantisation of a Scalar Field ...... 11 2.3.1 Poisson Brackets ...... 11 2.3.2 Quantisation ...... 11 2.3.3 The Hamiltonian for the Klein-Gordon field ...... 11 2.3.4 The Harmonic Oscillator in Quantum Mechanics ...... 12 2.3.5 Applying the Quantisation of the Harmonic Oscillator ...... 14 2.3.6 Infra-red and Ultra-violet Divergences ...... 15
3 The Lorentz & Poincar´egroups 16 3.1 Lorentz Transformations ...... 16 3.1.1 Defining a Lorentz Transformation ...... 16 3.1.2 Lorentz Transformations on Scalar Fields ...... 16 3.1.3 Introduction to the Lorentz Group ...... 17 3.2 Representations and Spinors ...... 19 3.2.1 Structure of the Lorentz Group ...... 19 3.2.2 Representations of the Lorentz Group ...... 21 3.2.3 Transformations of Spinors ...... 23 3.3 The Poincar´eGroup ...... 24 3.3.1 The Poincar´eGroup ...... 24 3.3.2 Casimirs of the Poincar´eGroup ...... 24
2 4 The Dirac Field 26 4.1 The Dirac Equation ...... 26 4.1.1 An introduction to Fermions ...... 26 4.1.2 The Dirac Equation: a first-order relativistic wave equation ...... 26 4.2 The Clifford Algebra ...... 27 4.2.1 The Dirac Basis ...... 27 4.2.2 The Weyl Basis ...... 28 4.3 Properties of Spinors and the Dirac Equation ...... 29 4.3.1 Lorentz Covariance of the Dirac Equation ...... 29 4.3.2 Dirac Bilinears ...... 30 4.3.3 The Dirac Lagrangian ...... 30 4.4 Canonical Quantisation of the Dirac field ...... 32 4.4.1 Plane Wave solutions ...... 32 4.4.2 Chirality & Helicity ...... 33 4.4.3 Formulae required for Quantisation ...... 33 4.4.4 Quantising the field ...... 35 4.4.5 The Hamiltonian ...... 36
5 Introducing Supersymmetry 39 5.1 The Supersymmetry Algebra ...... 39 5.1.1 What is Supersymmetry? ...... 39 5.1.2 The Supersymmetry Algebra ...... 39 5.1.3 Extended Supersymmetry ...... 40 5.1.4 Forming the Algebra ...... 40 5.1.5 Interpretation of the Algebra ...... 42 5.2 Supermultiplets ...... 43 5.2.1 What are Supermultiplets? ...... 43 5.2.2 Constructing Massive Supermultiplets ...... 44 5.2.3 Constructing Massless Supermultiplets ...... 45
6 The Free Wess-Zumino Model 47 6.1 Introducing the Model ...... 47 6.1.1 A Supersymmetric Lagrangian ...... 47 6.2 Modifying the Model ...... 49 6.2.1 Commuting SUSY transformations ...... 49 6.2.2 On-shell and Off-shell SUSY ...... 49 6.2.3 Adding an Auxiliary Field ...... 50 6.2.4 Commuting modified SUSY transformations ...... 51 6.3 Obtaining the SUSY algebra ...... 52 6.3.1 The Supercurrent ...... 52 6.3.2 The Supercharges ...... 53 6.3.3 Recovering the SUSY Algebra ...... 54
7 Conclusion & Further Discussion 56 7.1 Further Discussion ...... 56 7.2 Conclusion ...... 56
Appendix A Spinor Indices and Identities 58
References 60
3 Chapter 1
Introduction
1.1 Preamble 1.1.1 Introduction Within the last century, the concept of symmetry has become fundamental to constructing an all-encompassing theory of our universe. The most successful theory we have to date, the stan- dard model, is essentially a theory of symmetries and how these lead to physical phenomena. The model hinges on an intrinsic property of fundamental particles, spin. It splits particles into two categories: fermions, with half integer spin, which make up matter, and bosons, with whole integer spin, which characterise particles that ‘carry’ the fundamental forces. When working with the standard model, we utilise quantum field theory (QFT), which replaces the notion of particles with fields. The theory brings together special relativity and quantum mechanics. In this re- 1 port, we examine spin-0 and spin- 2 fields. We exclusively discuss free quantum field theory, and choose not to introduce interactions, which usually involves adding a potential term to our theory.
Given that QFT is constructed with relativity in mind, Lorentz transformations become crucial in understanding why we observe the particles we do. These transformations form the Lorentz group, the representations of which correspond to different types of fields: the trivial representation refers to a spin-0 (bosonic) scalar field, the spinorial representation refers to spin- 1 2 (fermionic) spinor fields, whilst higher spin representations correspond to gauge bosons and other such particles.
Supersymmetry (SUSY), is a relatively new theory that proposes a space-time symmetry (i.e. dependent on space-time co-ordinates) between fermionic and bosonic fields. The basic premise of the theory is the existence of supersymmetric generators, which turn a bosonic field into a fermionic field (and vice-versa). Thus, all physical particles must have ‘superpartners’, which are coupled together in ‘supermultiplets’. As of writing, the theory is severely lacking in experimen- tal evidence, but its mathematical elegance has provided reason enough for it to become a large research focus. In 1975, Haag, Sohnius and Lopuszanski proved that a supersymmetric theory was the only possible space-time symmetry extension to an interacting quantum field theory. Further to this, SUSY can be shown to ‘solve’ many of the problems with the standard model, such as the ‘fine-tuning problem’, which we unfortunately do not have time to review.
In this report, we focus on constructing a basic supersymmetric theory, the free massless
4 Wess-Zumino model, which consists of a single fermion and a single boson. We check that this transformation is a symmetry, and recover the associated supersymmetric algebra by applying results from QFT.
1.1.2 Conventions Throughout this report, we will be using index notation throughout with the space-time index µ = 0, 1, 2, 3 with Einstein’s summation convention implied. For a 4-vector aµ we have:
µ X µ a aµ = a aµ. (1.1) µ=0,1,2,3
Additionally, we choose to use natural units: those such that ¯h = c = 1. We will be exclusively using the Minkowski metric with the mostly-minus convention:
1 0 0 0 µν µν 0 −1 0 0 g = η = . (1.2) 0 0 −1 0 0 0 0 −1
We define our 4-gradient by
∂ ∂ ∂ ∂ ∂ ∂ = ∂ = , ∇ = , , , . (1.3) ∂xµ µ ∂t ∂t ∂x1 ∂x2 ∂x3
Using these conventions we write the 4-momentum, Pµ as
Pµ = (E, −p), (1.4) where p is the 3-momentum. Once a system is quantised (as discussed later) we obtain the 4-momentum operator, ∂ Pˆ = (E,ˆ −pˆ) = i , ∇ = i ∂ . (1.5) µ ∂t µ
5 Chapter 2
Field Theory
2.1 Classical Field Theory 2.1.1 Background Field theory replaces the concept of a single particle generalised co-ordinate to that of a contin- uum of particles, where we may have several particles at any point in space-time. We denote a field, φ(x), as a function of space-time co-ordinates. The main quantity of interest in classical mechanics is a functional named the action, S, defined as Z Z S = L dt = (pq˙ − H) dt, (2.1) where H is the Hamiltonian, H = T + V (the sum of kinetic and potential energies), L is the Lagrangian, given by L = T −V , and p and q are conjugate variables for generalised co-ordinates and momenta. Note the first half of this chapter follows closely to [1]. In field theory we opt to use the Lagrangian density, L = R L d3x, which we can write in terms of our fields. Hence we can write our action in terms of the Lagrangian density, L, as follows: Z 4 S = L (φ, ∂µφ) d x , (2.2)
∂φ where ∂µφ = ∂xµ and µ is the space-time index µ = 0, 1, 2, 3. Throughout this report, we will use the term Lagrangian to refer to the Lagrangian density. Though this is not formally correct this is done for ease given the Lagrangian density is the main quantity we will be dealing with.
2.1.2 Equations of Motion This subsection follows closely to [2]. We find the classical equations of motion for a system by applying the principle of stationary action. For this we consider an arbitrary volume in space- time, Ω, and vary our field φ(x) → φ0(x) = φ(x) + δφ(x) such that δφ = 0 on the volume surface, ∂Ω. Hence, ∂L ∂L δL = δφ + δ(∂µφ) (2.3) ∂φ ∂ (∂µφ) ∂L ∂L = δφ + ∂µδφ. (2.4) ∂φ ∂ (∂µφ)
6 This gives variation in the action, Z δS = d4x δL (2.5) Z Z 4 ∂L ∂L 4 ∂L ∂L = d x − ∂µ δφ + d x ∂µ δφ + ∂µδφ , (2.6) ∂φ ∂ (∂µφ) ∂ (∂µφ) ∂ (∂µφ) where we have inserted an extra term which has been both added and subtracted, so that we can now perform integration by parts: Z Z 4 ∂L ∂L 4 ∂L δS = d x − ∂µ δφ + d x ∂µ ∂µδφ . (2.7) ∂φ ∂ (∂µφ) ∂ (∂µφ)
The second integral vanishes by Gauss’ Divergence Theorem; it is just a surface/boundary term. By inferring the action is stationary we have δS = 0. Applying the fundamental theorem of calculus of variations we find the classical equations of motion for our field, the Euler-Lagrange equations: ∂L ∂L − ∂µ = 0. (2.8) ∂φ ∂ (∂µφ) The above form of φ(x) is a real scalar field, i.e. it has a single (real) value for every space-time co-ordinate. In general, fields may also be complex, and not necessarily scalar; we will discuss Lagrangians involving spinor fields later in this report. Note that if φ(x) was a complex field, there would be two separate equations of motion for the system, one for φ(x) and the other for its conjugate, φ∗(x).
2.1.3 Noether’s Theorem We continue to follow [1,2], breaking down calculations in the following derivations and examples. A symmetry of a system is defined as a set of transformations that leaves the action invariant. Noether’s (first) theorem states that for every symmetry there is an associated conserved quantity, usually referred to as the Noether current. To find this quantity, we apply a similar process to that of before, considering a variation of S under a continuous transformation of our field,
φ(x) → φ0(x) = φ(x) + α∆φ(x), (2.9) where α is an infinitesimal transformation parameter and ∆φ(x) is the shift in our field. Total derivatives in the Lagrangian leave nothing but a surface term in our integral for the action, so we can write the shift in the Lagrangian as
µ L(x) → L(x) + α∂µJ (x), (2.10) where we are free to choose J µ arbitrarily. We then can look at what our shift in L should be, given it is a function of φ(x) and ∂µφ(x):
∂L ∂L α∆L = (α∆φ) + ∂µ(α∆φ) (2.11) ∂φ ∂(∂µφ) ∂L ∂L ∂L = α∂µ ∆φ + α − ∂µ ∆φ. (2.12) ∂(∂µφ) ∂φ ∂(∂µφ) | {z } =0
7 The second term in the above equation vanishes by the Euler-Lagrange equations. Comparing terms,
∂L µ ∂µ ∆φ = ∂µJ (2.13) ∂(∂µφ) ∂L µ ⇒ ∂µ ∆φ − J = 0. (2.14) ∂(∂µφ)
µ Hence have a conserved current satisfying ∂µj (x) = 0, where
∂L jµ = ∆φ − J µ. (2.15) ∂(∂µφ)
There is an associated Noether charge found by integrating the zeroth component as follows: Z Q = d3x j0. (2.16)
We will use a similar process to that discussed in order to find our supercharges (supersymmetric generators) in later chapters.
2.1.4 An example of Noether’s Theorem The Lagrangian for a complex scalar field is given by 1 1 L = (∂ φ)(∂µφ)∗ − m2φφ∗. (2.17) 2 µ 2 We can apply Noether’s theorem using the following transformation as our symmetry:
φ → eiαφ. (2.18)
This is an example of a U(1) symmetry. As stated before, we consider our transformation in the form φ(x) → φ0(x) = φ(x) + α∆φ(x), where α is an infinitesimal parameter. This means we can write eiα ≈ 1 + iα, and so find that α∆φ = iαφ. (2.19) The conjugate field, φ∗, transforms very similarly, but with a negative sign from the conjugation. As we now have two independent fields, we must modify our earlier definition for the Noether current as summing over the fields: Using the definition of our conserved current, we find
X ∂L jµ = ∆Φ − J µ ∂(∂µΦ) Φ=φ,φ∗ (2.20) = (∂µφ)∗(iφ) + (−iφ∗)(∂µφ) = i((∂µφ∗)φ − φ∗(∂µφ)), where we have chosen J µ = 0. This particular current is important in field theory when dis- 1 µν cussing the Lagrangian for electromagnetic fields, L = − 4 Fµν F .
8 2.1.5 On-shell and Off-shell When using the term on-shell we mean to say that the classical equations of motion, the Euler- Lagrange equations, are obeyed. If this is not the case, we have an off-shell configuration. In a relativistic quantum system, where virtual particles are free to be created and destroyed, the term on-shell more specifically refers to the virtual particles satisfiying Einstein’s mass-energy relation, p2 + m2 = E2, (2.21) where p is the 3-momentum of the particle. Noether’s Theorem is an example of an on-shell result: as shown in finding an expression for the Noether current jµ, we use the Euler-Lagrange equations to infer the form of the current stated.
2.2 Moving to Quantum Field Theory 2.2.1 The Klein-Gordon Equation We can find the equations of motion for a real scalar field by quantising a system (to be discussed in further depth shortly). We use the standard quantisation, replacing physical observables with operators, written in natural units as: p → pˆ = −i∇, (2.22) ∂ E → Eˆ = i . (2.23) ∂t We use the standard mass-energy relation given above, and replace our observables with opera- tors, noting that we must introduce a field given that we have quantised the system: pˆ2 + m2 φ = Eˆ2φ ∂ 2 (−i∇)2 + m2 φ = i φ, ∂t (2.24) ∂2 ⇒ − ∇2 + m2 φ = 0. ∂t2 Writing this in index notation we have
µ 2 (∂ ∂µ + m )φ = 0. (2.25) This is the Klein-Gordon equation: the simplest relativistic wave equation, which serves as the equation of motion for the field φ [2]. We can find the same expression by using the Euler- Lagrange equations, noting that the Lagrangian for a massive real scalar field is given by: 1 1 L = (∂ φ)(∂µφ) − m2φ2. (2.26) 2 µ 2 We then find ∂L = −m2φ, (2.27) ∂φ ∂L = ∂µφ, (2.28) ∂(∂µφ) which, when placed in the Euler-Lagrange equations, recover the Klein-Gordon equation, 2 µ µ 2 −m φ − ∂µ(∂ φ) = (∂ ∂µ + m )φ = 0. (2.29)
9 2.2.2 Conjugate Momentum and Hamilton’s Equations We continue to follow sources [1,2] closely for the rest of this section. In classical mechanics, we have an alternative description to the Lagrangian formalism for the motion of a system. This is the Hamiltonian formalism, where the Hamiltonian is obtained by performing a Legendre trans- form on our Lagrangian: H = pq˙ − L. Here we choose to use canonically conjugate momentum, given by ∂L p = . (2.30) ∂q˙ Instead of a single second order differential equation (the Euler-Lagrange equations of motion), we can describe a system using two first order differential equations, called Hamilton’s equations. These are: ∂H ∂H p˙ = − , q˙ = . (2.31) ∂q ∂p Analogous to classical mechanics, in field theory we define the conjugate momentum by
∂L(φ, ∂µφ) ∂L(φ, ∂µφ) π(x) = = = ∂0φ(x). (2.32) ∂φ˙ ∂(∂0φ) From this we can define our field theory Hamiltonian by once again applying a Legendre trans- form: Z Z H = d3x H(φ, π) = d3x (πφ˙ − L), (2.33) where H is the Hamiltonian density.
2.2.3 The Stress-Energy Tensor An important symmetry in nature is that of space-time translations; translating our co-ordinates should not change the action. More specifically, the transformation can be written as xµ → xµ − aµ. (2.34) This means the field transforms as µ φ(x) → φ(x + a) = φ(x) + a ∂µφ(x) (2.35) and similarly our Lagrangian as ν µ L → L + a ∂µ(δν L). (2.36) µ Comparing this the general form of our transformations, L(x) → L(x) + a∂µJ (x), we identify µ µ (J (x))ν = L δν . (2.37) µ So we have an associated conserved current, which we call the Stress-Energy Tensor, Tν , given by
µ ∂L µ Tν = ∂ν − L δν . (2.38) ∂(∂µφ) Integrating over the various parts of this tensor gives four different conserved charges. These are the Hamiltonian (which is conserved for translations in time), and the components of physical 3-momentum (conserved for translations in space): Z Z H = T 00 d3x = H d3x, (2.39) Z Z i 0i 3 3 P = T d x = − π∂iφ d x. (2.40)
10 2.3 Canonical Quantisation of a Scalar Field 2.3.1 Poisson Brackets In classical mechanics, we can write the classical equations of motion in the Hamiltonian formu- lation by using Poisson Brackets. Using canonical co-ordinates we can define the Poisson Bracket of two functions, f(qi, pi) and g(qi, pi) as follows: X ∂f ∂g ∂f ∂g {f(q , p ), g(q , p )} = − . (2.41) i i i i ∂q ∂p ∂p ∂q i i i i i One reason for introducing Poisson Brackets is that they allow us to write a symmetric form of Hamilton’s equations: q˙i = {qi,H(qi, pi)}, p˙i = {pi,H(qi, pi)}. (2.42) We have the following important Poisson Bracket relations:
{qi, pj} = δij, (2.43)
{qi, qj} = {pi, pj} = 0. (2.44)
2.3.2 Quantisation In quantum mechanics we quantise a system by promoting our canonical variables (and other physical observables) to operators which act on a systems state. To signify this we place a hat on the operator, i.e.q ˆi. The general procedure involves introducing commutators, defined as [A, B] = AB − BA, which take the same form as their Poisson bracket counterparts but with a factor of i¯h. In natural units we have:
[ˆqi, pˆj] = i δij, (2.45)
[ˆqi, qˆj] = [ˆpi, pˆj] = 0. (2.46) where as previously mentioned our 3-momentum becomes a differential operator:
p → pˆ = −i∇. (2.47)
In quantum field theory we undertake a similar approach, utilising the conjugate momentum found earlier to write our commutators as h i φˆ(x), πˆ(y) = i δ(3)(x − y), (2.48) h i h i φˆ(x), φˆ(y) = πˆ(x), πˆ(y) = 0. (2.49)
Note how we now have a Dirac delta function, given we have a continuous system as opposed to a discrete one. The above equations are known as equal time commutation relations. For ease of notation, we will drop the hat from our operators throughout this report.
2.3.3 The Hamiltonian for the Klein-Gordon field For this and the subsequent subsection, we follow [1] exclusively, and carry out all calculations explicitly. The Lagrangian for a Klein-Gordon field can be written as 1 1 1 1 1 L = (∂ φ)2 − m2φ2 = φ˙2 − (∇φ)2 − m2φ2. (2.50) 2 µ 2 2 2 2
11 Using our earlier definition for the associated Hamiltonian is we find that Z H = d3x (πφ˙ − L) Z 3 ˙ 1 ˙2 1 2 1 2 = d x πφ − 2 φ − 2 (∇φ) − 2 m (2.51) Z 3 1 2 1 2 1 2 2 = d x 2 π + 2 (∇φ) + 2 m φ , where π = φ˙. Before going further, it is important to mention that the above field theory commutation relations, as well as our Hamiltonian, clearly treat spatial and time co-ordinates differently. Hence the Hamiltonian formulation is not ‘Lorentz invariant’, as well be discussed in greater depth shortly. However, this formulation makes quantisation far easier. Note in our commutation relations and thereafter we assume the Schr¨odingerpicture, in which the operators themselves are time-independent. We can use a Fourier expansion for our field φ(x, t): Z d3p φ(x, t) = eip.x φ(p, t), (2.52) (2π)3 where φ is real, and so φ∗(p) = φ(−p). Using this decomposition, the Klein-Gordon equation becomes: ∂2 + (p2 + m2) φ(p) = 0. (2.53) ∂t2 This equation is of the same form as the harmonic oscillator, where the natural frequency of the oscillator, ωp, is given by p 2 2 ωp = p + m . (2.54) From this we can say the general solution to the above equation is a linear superposition of oscillators with different frequencies, given by different momenta.
2.3.4 The Harmonic Oscillator in Quantum Mechanics The general Hamiltonian for SHO in quantum mechanics is given by
1 2 1 2 2 H = 2 p + 2 ω q , (2.55) where q and p obey commutation relations as previously outlined. We introduce ladder operators, a and a†, which as will be shown explicitly soon, act on a state to produce another state of lower or higher energy [3]. These are given by:
rω i a = q + √ p, (2.56) 2 2ω rω i a† = q − √ p, (2.57) 2 2ω which means we can write our canonical co-ordinates as 1 q = √ a + a† , (2.58) 2ω rω p = −i a − a† . (2.59) 2
12 Using the fact [q, p] = i, we can then find the commutation relation for a and a†:
1 rω [q, p] = i = √ a + a† , −i a − a† 2ω 2 i = − a + a†, a − a† 2 i † † † † (2.60) = − [a, a] − a, a + a , a − a , a 2 |{z} | {z } =0 =0 i = − −2 a, a† 2 = i a, a† . Hence we have the commutator a, a† = 1. (2.61) Our Hamiltonian becomes: 2 1 rω 1 1 2 H = −i a − a† + ω2 √ a + a† 2 2 2 2ω ω 2 ω 2 = − (a2 + a† − aa† − a†a) + (a2 + a† + aa† + a†a) 4 4 (2.62) ω ω = (aa† + a†a) = (a, a† + 2a†a) 2 2 † 1 = ω(a a + 2 ). Using these results we can find the commutation relations of the ladder operators with the Hamiltonian:
[H, a†] = ωa†, (2.63) [H, a] = −ωa. (2.64)
The rest of this subsection follows closely to [3]. We can use these commutation relations to act on a state with energy E, where H |Ei = E |Ei. We find the energy eigenstates of states that have had a ladder operator applied on them as follows:
Ha† |Ei = (a†H + ωa†) |Ei = (E + ω)a† |Ei , (2.65) Ha |Ei = (aH − ωa) |Ei = (E − ω)a |Ei . (2.66)
If the energy is bounded from below, such that we have a ground state |0i, then the ladder operator a must satisfy a |0i = 0. Hence when the Hamiltonian acts on the ground state we have
1 H |0i = 2 ω |0i . (2.67) If we define a state in the nth energy level as
|ni = (a†)n |0i , (2.68) then we find applying the Hamiltonian to such a state gives
1 H |ni = (n + 2 )ω |ni . (2.69) 1 + Hence we see the spectrum of the SHO Hamiltonian can be given by (n + 2 ) where n ∈ Z .
13 2.3.5 Applying the Quantisation of the Harmonic Oscillator Moving back to the free scalar field, we make analogy to our definitions of q and p involving ladder operators by defining our field and conjugate momentum. The following calculations follow [1,3] but we show the steps in far greater detail. Once again expanding over all the different oscillator modes, Z d3p 1 φ(x) = a eip.x + a† e−ip.x , (2.70) 3 p p p (2π) 2ωp Z d3p rω π(x) = (−i) p a eip.x − a† e−ip.x , (2.71) (2π)3 2 p p † where ap and ap are our creation and annihilation operators, dependent on the 3-momentum index p we are summing over. The next part of the analogy involves finding commutators for the creation and annihilation operators. We use the ansatz † 3 (3) ap, aq = (2π) δ (p − q), (2.72) † † [ap, aq] = ap, aq = 0, (2.73) and then try to recover the known commutation relation for our fields: Z 3 3 r d p d q −i ωq ip.x † −ip.x iq.y † −iq.y [φ(x), π(y)] = 3 3 ape + ape , aqe − aqe (2π) (2π) 2 ωp Z 3 3 r d p d q −i ωq † i(p.x−q.y) † i(−p.x+q.y) = 6 ap, −aq e + ap, aq e (2π) 2 ωp Z 3 3 r d p d q −i ωq (3) i(p.x−q.y) (3) i(−p.x+q.y) (2.74) = 3 −δ (p − q)e − δ (q − p)e (2π) 2 ωp Z d3p i Z d3p = eip.(x−y) + eip.(y−x) = i eip.(x−y) (2π)3 2 (2π)3 = i δ(3)(x − y). Hence our ansatz is correct. Our Hamiltonian then becomes Z 3 1 2 1 2 1 2 2 H = d x 2 π + 2 (∇φ) + 2 m φ (2.75) 3 3 √ 1 Z Z d p d q ωpωq = d3x − a eip.x − a† e−ip.x a eiq.x − a† e−iq.x 2 (2π)3 (2π)3 2 p p q q 1 ip.x † −ip.x iq.x † −iq.x + √ i p ape − i p ape i q aqe − i q aqe (2.76) 2 ωpωq 2 m ip.x † −ip.x iq.x † −iq.x + √ ape + ape aqe + aqe . 2 ωpωq
R d3x i(k−k0).x (3) 0 We then integrate over x making use of the identity (2π)3 e = δ (k − k ). After the delta functions have been applied we have: Z 3 d p 1 n 2 † † † † H = 3 − ωp apa−p − apap − apap + apa−p (2π) 4 ωp † † † † (2.77) − p. − p apa−p − p.p apap − p.p apap + p. − p apa−p
2 † † † † o + m apa−p + apap + apap + apa−p .
14 2 2 2 Using ωp = p + m many terms then cancel:
Z d3p −ω Z d3p ω H = p −2 a† a − 2 a a† = p a† a + a a† (2π)3 4 p p p p (2π)3 2 p p p p Z d3p = ω a† a + 1 a , a† (2.78) (2π)3 p p p 2 p p Z d3p Z d3p ω = ω a† a + p δ(3)(0), (2π)3 p p p (2π)3 2 where we have use the commutation relations of creation and annihilation operators.
2.3.6 Infra-red and Ultra-violet Divergences This discussion follows [3]. Note that second integral above has noticeable convergence issues. Firstly, the integral diverges because we are integrating over an infinitely large space. The integral produces an infinity as we are looking to find the energy for an infinite volume; a more poignant question is to look for the energy density for a finite volume. As outlined in [3], we can take define our volume, V , as a box of length L, and say that
Z L/2 Z L/2 3 (3) 3 ip.x 3 (2π) δ (0) = lim d x e p=0 = lim d x = V. (2.79) L→∞ −L/2 L→∞ −L/2
So our energy density, ε0, can be written as:
E Z d3p 1 ε = 0 = ω . (2.80) 0 V (2π)3 2 p
The divergence due to the delta function is known as an infra-red divergence. However, there are still problems with the above integral given for energy density. We are integrating over infinitely large momentum modes, with the ωp term tending to infinity as momentum does. Hence our integral diverges. This divergence is an ultra-violet divergence. This is much more tricky to remove, with a maximum limit for our momentum, Λ, needed for the integral to converge. Instead, we choose to remove this term for our expression for the Hamiltonian, effectively setting our vacuum ground state to zero. As we are only interested in relative energies, this is an acceptable thing to do. Hence the Hamiltonian becomes
Z d3p H = ω a† a , (2.81) (2π)3 p p p which gives H |0i = 0 |0i as outlined.
15 Chapter 3
The Lorentz & Poincar´egroups
3.1 Lorentz Transformations 3.1.1 Defining a Lorentz Transformation A Lorentz transformation is given by an O(3, 1) matrix which acts on space-time co-ordinates such that µ µ µ ν x → x˜ = Λ ν x , (3.1)
µ ∂x˜µ where we define our Lorentz transformation as Λ ν = ∂xν , which preserves the inner product of µ ν µ our co-ordinates, x · x = x gµν x = x xµ, such that: µ σ ν ρ σ ρ (Λ σx ) gµν (Λ ρx ) = x gσρ x . (3.2) We can then write this condition more compactly as µ ν σρ µν Λ σ Λ ρ g = g . (3.3) This section follows closely to [1, 3, 6]. This implies that we have some measure of space-time, µ usually denoted ∆ = x xµ, that is left invariant. The description O(3, 1) refers to the matrices belonging to the orthogonal group, implying they are real, have determinant ±1 with inverse matrix the same as their transpose matrix. Further to this we restrict our transformations to those which are proper and orthochronous. By ‘proper’ we mean these matrices which have determinant +1. By orthochronous, the transformations preserve the direction of time, i.e. 0 ↑ Λ 0 > 0 [4]. With this restriction, we write our new subgroup as SO+(3, 1) , referring to a special orthogonal group with these two conditions applied.
3.1.2 Lorentz Transformations on Scalar Fields Continuing on the same path set in the previous chapter, we would like to look at how scalar fields transform under a Lorentz transformation. The simplest way we can write is given by µ ˜ µ −1 µ ν φ(x ) → φ(x ) = φ((Λ ) ν x ). (3.4) Our intuition in writing this comes from the idea that the transformed field takes the same value at a point that it would have taken before a Lorentz transformation was applied [5]. Note that from the condition given previously, our Lorentz transformations obey the identity −1 ρ −1 σ µν ρσ (Λ ) µ (Λ ) ν g = g , (3.5)
16 which we can use to check Lorentz invariance. A Lorentz-invariant quantity is one that remains unchanged under a Lorentz transformation (a combination of rotations and Lorentz boosts). For 1 2 1 2 2 example, in the Klein-Gordon field Lagrangian, L = 2 (∂µφ) − 2 m φ , the second (mass) term will remain unchanged under a transformation. The first (kinetic) term, transforms as
2 µν ˜ ˜ (∂µφ (x)) → g ∂µφ (x) ∂ν φ (x) µν −1 −1 = g ∂µ φ Λ x ∂ν φ Λ x ρ ρ = gµν Λ−1 x Λ−1 ∂ φ Λ−1 ∂ φ µ ρ ν σ ; (3.6) µν −1ρ −1ρ −1 = g Λ µ Λ ν Λ x (∂ρφ)(∂σφ) ρσ −1 = g Λ x (∂ρφ)(∂σφ) −1 2 = Λ x (∂µφ (x)) .
Hence we can say the Lagrangian transforms as a scalar,
L(φ(x)) → L(φ(Λ−1 x)). (3.7)
The above calculation is given in [1], but shown more explicitly here. Integrating over all space- time, we find the action unchanged under this transformation and so we say the action is Lorentz invariant. Similarly, we can show the Lorentz invariance of the Klein-Gordon equation using the same technique: