World Scientific Lecture Notes in Physics — Vol. 80 INTRODUCTION TO SUPERSYMMETRY (Second Edition) Harald J W Miuller-Kirsten University of Kaiserslautern, Germany Armin Wiedemann Baden-Wurttemberg Cooperative State University Mannheim, Germany WP world Scientific NEW OIEKSEY © Fontnial cinPublished by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. World Scientific Lecture Notes in Physics — Vol. 80 INTRODUCTION TO SUPERSYMMETRY (2nd Edition) Copyright © 2010 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or byany means, electronic or ‘mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher. For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher. ISBN-13 978-981-4293-41-9 ISBN-10 981-4293-41-5 ISBN-13 978-981-4293-42-6 (pbk) ISBN-10 981-4293-42-3 (pbk) Printed in Singapore by B & JO EnterprisePreface to the Second Edition The positive response to our text published two decades ago and its quick sell-out encouraged us after careful consideration to produce this second and IsTpX-typed version. We observed that our text was used or recommended in courses at several universities. Naturally, the most positive response (partic- ularly from the US) came from those for whom the text was written, namely students and others without prior knowledge of supersymmetry. We thank our readers for this feedback and other comments. This success assured us that indeed there was a definite demand for a text like our’s with explicit calculational details. In this second edition we have corrected writing errors we became aware of. We have also made minor changes here and there, and we cite more literature. In the course of the last 20 years since the publication of the First Edition, a vast amount of literature on supersymmetry and related topics has been published. Various non-technical books on supersymmetry, e.g. D. Hooper (58], G. Kane [62], L. Randall [95], and the books by B. Greene [52,53] are now available. There are several introductory texts on supersymmetry, e.g. LJ.A. Aitchinson {1,2], A. Bilal [15], M. Drees [32], N. Polonsky [89], 8.P. Misra [73], U. Lindstrém [69], J.D. Lykken [70], S.P. Martin [71]. Books on superstrings generally include brief introductions to supersymmetry, see e.g. the monographs by M.B. Green, J.H. Schwarz and B. Witten [51], M. Kaku [61] and J. Polchinski [88]. Phenomenological implications and astrophysical and cosmological aspects of supersymmetry — subjects which we do not cover in this text — can be found in M. Dine [28], K.A. Olive [80] and S.P. Martin [71]. The Minimal Supersymmetric Standard Model (MSSM) — i.e. the minimal extension to the Standard Model of particle physics that realizes supersymmetry with a strong emphasis on phenomenological aspects — is discussed in I.J.A. Aitchinson [1] and in the review article by S.P. Martin [71]. This subject is also treated in the monograph by S. Weinberg [118], Chaptervi PREFACE TO THE SECOND EDITION 28.4. Non-perturbative methods of supersymmetry are discussed in A. Wipf [129], J. Terning [111] and M. Bianchi et al. [13]. More advanced topics such as the Seiberg-Witten duality — an important development of non- perturbative methods in supersymmetric Yang-Mills Theory — is presented in the pedagogical reviews of L. Alvarez-Gaumé and S.L. Hassan [5], A. Bilal [14], and W. Lerche [68]. The book by D.I. Olive and P.C. West [79] is also devoted to this topic. Recently the books of $. Duplij, W. Siegel, and J. Bagger [33], J. Terning [112], M. Dine [29], and P. Binetruy and K. Hentschel [16] appeared, which cover the developments of the last two decades in supersymmetry. A few years ago, the discovery of supersymmetry celebrated its 30‘ an- niversary. The exceptional history of supersymmetry is discussed in the book by G. Kane and M. Shifman [63], in particular we recommend the compre- hensive contribution of R. Di Stefano [30]. See also the recollections of J. Wess [122] and the overview of D. Olive and P. West: in 79]. One of us (A.W.) thanks Mees de Roo, University of Groningen, for numerous suggestions and his encouragement. Finally we thank the editors of World Scientific for their support and cooperation. Summer 2009 H.J.W. Miiller-Kirsten University of Kaiserslautern A. Wiedemann Baden-Wiirttemberg Cooperative State University MannheimPreface to the First Edition This text is a detailed version of material presented by both of us in semi- nars and lectures in the theory group of this department. Except for parts of Chapters 9 and 10 the material has also been covered in a series of sem- inars by one of us (M.-K.) in the Department of Physics of the University of Adelaide, Adelaide, Australia, in August and September 1985 and in the Department of Physics of Shanxi University, Taiyuan, China, in March and April 1987. The interest and criticism of the audience at these departments and, in particular, the support and enthusiasm of Prof. A.W. Thomas (Ade- laide) and Professor Zhang Jianzu (Taiyuan) are gratefully acknowledged. The text was compiled with the belief that the majority of potential readers is more interested in actually using or applying supersymmetry in some model theory than in painstakingly rediscovering the results of others for themselves. It seemed plausible, therefore, to revise various relevant concepts and in particular, to include the proof or verification of almost every formula. In this way the reader can select the problems he wants to tackle himself, compare his solutions with the calculations given here, and thus gain the confidence in his own calculations which he needs for his discussions of supersymmetry in other contexts. It has been our experience that except for the last two chapters, the material presented here can be covered in a one- semester course for graduate or post graduate students with some knowledge of field theory. In compiling this text we have, of course, used previous reviews. The choice of our sequence of topics was motivated by the lecture notes of F. Legovini [67]. Standard texts which we have consulted are the monograph by J. Wess and J. Bagger [123] and the review by P. Fayet and S. Ferrara [35]. For the detailed treatment of the on-shell Wess-Zumino model we consulted the lecture notes of M. de Roo [25]. In the text we do not discuss any experimental signatures of supersymmetry. For an introduction into this topic we refer to articles by H.E. Haber and G.L. Kane [55,56]; further details can be found in the Proceedings of the Thirteenth SLAC Summer viiviii PREFACE TO THE FIRST EDITION Institute on Particle Physics [92] and in the reviews by P. Nilles [77] and N. Dragon, U. Ellwanger and M. Schmidt [31]. As further general references we refer to the nontechnical review by J. Wess [121], to a very brief review of topics covered here by B.A. Campbell and G. Fogleman [20] and to the lectures by J. Wess [120], S. Ferrara [37] and E. Witten [132]. A more advanced text is the book by S.J. Gates, M.T. Grisaru, M. Rotek and W. Siegel [46]. The very readable review by M. Sohnius [109] appeared after completion of the first draft of our text. Meanwhile several other texts have been published, each, however, with its emphasis in a different direction. We refer here to the books by P. West [127], P.P. Srivastava [110] and P.G.O. Freund [44]. For more specific topics we refer to the article by A. Salam and J. Strathdee [101], to the Proceedings of the 28th Scottish Universities Summer School in Physics [90], and to the Proceedings of the NATO Advanced Study Institute on Supersymmetry [91]. All considerations of this text refer to a four-dimensional Minkowski space. For the basic technicalities in the context of supersymmetric quantum mechanics we refer to the work of F. Cooper and B. Freedman [24], whereas those of two-dimensional field theories can be found in reference [74]. Summer 1987 H.J.W. Miiller-Kirsten and A. Wiedemann University of KaiserslauternContents Preface to the Second Edition v Preface to the First Edition vii Introduction 1 1 Lorentz and Poincaré Group, SL(2,C), Dirac and Majorana Spinors t 1.1 The Lorentz Group. ... 0.0.0.0 7 1.2 The Poincaré Group .......---- 21 1.3 SL(2,C), Dotted and Undotted Indices 32 1.3.1 Spinor Algebra... 2.0... 0 32 1.3.2 Calculations with Spinors 48 1.3.3 Connection between SL(2,€) and Li... 2... 53 1.3.4 The Fierz-Reordering Formula .........--+-- 64 1.3.5 Further Calculations with Spinors 65 1.3.6 Higher Order Weyl Spinors and their Representations 78 1.4 Dirac and Majorana Spinors... 2... ee ee ee 83 1.4.1 The Weyl Basis or Chiral Representations ...... . 85 1.4.2 The Canonical Basis or Dirac Representation .... - 92 1.4.3 The Majorana Representation 1.4.4 Charge Conjugation, Dirac and Weyl Representations 101 1.4.5 Majorana Spinors .......----- 20+ ee eee 109 1.4.6 Calculations with Dirac Spinors . . 112 1.4.7. Calculations with Majorana Spinors 114 2 No-Go Theorems and Graded Lie Algebras 117 2.1 The Theorems of Coleman-Mandula and Haag, Lopuszaiski, SOMA eerste eaen ie aarece seer ree ae aceasta 117 2.1.1 The Theorem of Coleman-Mandula ecole 2.1.2 The Theorem of Haag, Lopuszariski and Sohnius .. . 119 ixCONTENTS 2.2 Graded Lie Algebras 2.2.1 Lie Algebras 2.2.2 Graded Algebras 2.2.3 Graded Lie Algebras... . 2.3 The Graded Lie Algebra of SU(2,C) . 2.4 Zz Graded Lie Algebras = 25) Graded Maries The Supersymmetric Extension of the Poincaré Algebra 3.1 Four-Component Dirac Formulation 3.2 Two-Component Weyl Formulation Representations of the Super—Poincaré Algebra 4.1 Casimir Operators 4.2 Classification of Irreducible Representations . . . . 4.2.1 N =1 Supersymmetry 4.2.2 N > 1 Supersymmetry The Wess—Zumino Model 5.1 The Lagrangian and the Equations of Motion . . . 5.2. Symmetries . 5.3 Plane Wave Expansions 5.4 Projection Operators... .. . 5.5 Anticommutation Relations . . . 5.6 The Energy-Momentum Operator of the Wess~Zumino Model 5.6.1 The Hamilton Operator . . 5.6.2. The Three-Momentum P; . . 5.7. Infinitesimal Supersymmetry Transformations ......... Superspace Formalism and Superfields 6.1 Superspace 6.2 Grassmann Differentiation . . ae 6.3 Supersymmetry Transformations in the Weyl oie 6.3.1 Finite Supersymmetry Transformations ........ 6.3.2 Infinitesimal Supersymmetry Transformations and Dif- ferential Operator Representations of the Generators . 6.4 Consistency with the Majorana Formalism . . . 6.5 Covariant Derivatives 6.6 Projection Operators... . 6.7 Constraints 6.8 Transformations of Component Fields 121 121 123 123 125 131 139 147 147 161 163 163 172 172 180 185 185 187 194 205 209 222 + 225 232 235 243 . 243 246 250 250Contents xi 7 Constrained Superfields and Supermultiplets 287 7.1 Chiral Superfields ReeeaeO te 7.2 Vector Superfields, Generalized Gauge Transformatio: . 300 7.3 The Supersymmetric Field Strength .............- 306 8 Supersymmetric Lagrangians 317 Si] Grasmann inteeruon 317 8.2 Lagrangians and Actions . 323, 8.2.1 Construction of Lagrangians from Scalar Superfields . 323 8.2.2 Construction of Lagrangians from Vector Superfields . 332 i223: Peerage eae 340 9 Spontaneous Breaking of Supersymmetry 343 9.1 The Superpotential. 2.0... 000. e eee eee 343 9.2 Projection Technique... ....- 0... 0+ eee eee 347 9.3 Spontaneous Symmetry Breaking . 364 9.3.1 The Goldstone Theorem... .. . 367 9.3.2 Remarks on the Wess-Zumino Model ........- 371 9.4 The O’Raifeartaigh Model... ....--- + eevee 372 9.4.1 Spontaneous Breaking of Supersymmetry .. . 372 9.4.2 The Mass Spectrum of the O’Raifeartaigh Model . eles 10 Supersymmetric Gauge Theories 387 10.1 Minimal Coupling .. 2... 1.1 eee ee ee 387 395, 400 412 10.2 Super Quantum Electrodynamics ..... . 10.3 The Fayet~Iliopoulos Model . 10.4 Supersymmetric Non-Abelian Gauge Theory Bibliography 423 Index 433Introduction Symmetries are of fundamental importance in the description of physical phenomena. In the realm of particle physics symmetries are believed to permit ultimately a classification of all observed particles. A fundamental symmetry of particle physics, which has been firmly established both theoret- ically and experimentally is that of the Poincaré group, i.e. of rotations and translations in four-dimensional Minkowski space. Besides this fundamental symmetry there are other so-called internal symmetries (such as the symme- try of the SU(3) flavor group) which have also been firmly established over the last few decades, although their manifestation in Nature is not exact. As is well known, the consistent search for more fundamental symmetries led to the development of non-Abelian gauge theories and the spectacular experimental confirmation of several predictions of the latter in recent years. In the course of time several attempts have been made to unify the space- time symmetry of the Poincaré group with the symmetry of some internal group ((22], [81], [82], [83]). Such attempts have, however, been shown to be futile if the theory, which necessarily has to be a quantum field theory, is expected to satisfy certain basic requirements. In fact, the so-called “no- go”-theorem of S. Coleman and J. Mandula [23] shows that if one makes the plausible assumptions of locality, causality, positivity of energy and finiteness of the number of particles (and one more technical assumption) the invari- ance group of the theory can at best be the direct product of the Poincaré group and a compact internal group, and this therefore does not offer a genuine unification of one group with the other. The generators of the Poincaré group satisfy well known commutation relations and Noether’s theorem relates these to conserved currents. In their turn these conserved currents are functions of relativistic fields. The commu- tation relations of the field operators which quantize these fields are therefore directly related to those of the generators. It was realized by J. Wess and B. Zumino [124], [125] that if one allows also anticommutation relations of generators of supersymmetry transformations which transform bosons into fermions and vice versa, then the unification of the spacetime symmetries 12 INTRODUCTION of the Poincaré group with this internal symmetry can be achieved. The formal proof of this discovery, i.e. the proof that anticommuting generators which respect the other assumptions of the theorem of S. Coleman and J. Mandula [23] do exist, was established by R. Haag, J.T. Lopuszariski and MF. Sohnius [54]. Supersymmetry thus arises as a symmetry which combines bosons and fermions in the same representation or multiplet of the enlarged group which encompasses both the transformations of the Poincaré group and the appro- priate supersymmetry transformations. Thus every bosonic particle must have a fermionic partner and vice versa. In view of the fact that such a spectrum of particles is not compatible with observation, supersymmetry must be badly broken at the level of presently available energies. Clearly only experimental observation can decide whether supersymmetry is indeed inherent in Nature. It can be argued that one of the most immediate ways to observe evidence of supersymmetry is to see if there is a missing energy and momentum in the final ete~ spectrum of the reaction ete ay SHE Sette +F7tF¥ where é*, é~ and ¥ are the supersymmetric partners of e+,e~ and y respec- tively. If there is such a missing energy and momentum it could be that carried away by the neutral photino ¥. Charged supersymmetry particles at energies presently available would have been detected long ago. Since supersymmetry must be broken, the photinos ¥ would not be massless. However, supersymmetry does not only open the possibility of a much more complex spectrum of particles than heretofore envisaged; supersym- metry also has some intriguing theoretical consequences which could make it a desirable theory. It is well kown that a realistic quantum field theory in the traditional is plagued by the problem of ultraviolet divergences and the consequent necessity of renormalization. Supersymmetry, however, provides a mechanism for the cancellation of such divergences in view of the same number of bosonic and fermionic degrees of freedom in each particle multiplet. Clearly, such a built-in cancellation of divergent terms is a highly desirable feature of a quantum field theory. In Chapter 1 we begin with a recapitulation of basic aspects of the Lorentz group, including a discussion of Casimir operators and the classification of representations in terms of their eigenvalues. We then consider the group SL(2,€) and its basic representations, i.e. the self-representation and the complex conjugate self-representation. The elements of the appropriate rep- resentation spaces are the undotted and dotted Weyl spinors. In view of the importance of Wey] spinors throughout the entire text, we consider these hereIntroduction 3 in more detail than is generally done in the literature. We then introduce the concept of Grassmann numbers and perform some basic manipulations involving Weyl spinors, thereby deriving a number of useful formulas. In the subsequent section the connection between the special linear group SL(2,C) and the proper orthochronous Lorentz group is established. It is then nat- ural to discuss four-component Dirac spinors and the Weyl representation. The connection with two-component Weyl spinors is obtained by introducing four-component Majorana spinors. Then again various formulas are derived which are useful in later calculations. Chapter 2 begins with a discussion of the “no-go” theorems of Coleman and Mandula [23] and Haag, Lopuszariski and Sohnius [54]. The latter leads to a consideration of graded Lie algebras which we approach in successive steps by defining first the characteristics of a Lie algebra, then those of a graded algebra and finally those of a graded Lie algebra, i.e. the properties of grading, supersymmetrization and generalized Jacobi identities. As an example we construct the graded Lie algebra of the algebra su(2,C). The final section of Chapter 2 deals with graded matrices and their properties. The following Chapter 3 deals with the grading, i.e. supersymmetrization of the Poincaré algebra. We demonstrate explicitly that for the grading chosen all possible Jacobi identities are satisfied. This turns out to be a crucial point of consistency of a grading. Having established the algebra of the Super~Poincaré group with the fermionic generators in the Dirac four- component form, we then decompose it into the appropriate relations of the two-component Weyl formalism. In Chapter 4 we use the method of Casimir operators to classify the irre- ducible representations of the Super—Poincaré algebra, and it is shown that supersymmetry implies an equal number of bosonic and fermionic degrees of freedom. Chapter 5 deals with the most immediate field theoretical realization of the Super—Poincaré algebra, the Wess-Zumino model, which is a field theory involving a scalar field, a pseudoscalar field and one spinor field, all with the same mass. We demonstrate by explicit calculation that the spinor charges of the theory, considered as linear operators in Fock space, satisfy the commutation and anticommutation relations of the Super—Poincaré algebra. In Chapter 6 we introduce the concepts of superspace and superfields, and define differentiation with respect to Grassmann numbers. Then three differ- ent but related operators are constructed which describe three different but equivalent actions of the supersymmetry group on functions in superspace. These operators define three different types of superfields. By considering infinitesimal supersymmetry transformations we obtain the corresponding three differential operator representations of the fermionic generators of the4 INTRODUCTION Super-Poincaré group. Then covariant derivatives are introduced as a pre- requisite for the construction of manifestly supersymmetric action integrals. These covariant derivatives also permit the definition of projection operators. The search for irreducible representations of the Super-Poincaré algebra then becomes a search for solutions of constraint equations expressed in terms of these projection operators. The final section of Chapter 6 is devoted to the derivation of the explicit supersymmetry transformations of the component fields of the supermultiplet. In this context it is seen that the highest or- der component field always transforms into a total Minkowski derivative and thus is a candidate for a supersymmetric Lagrangian density. In Chapter 7 we begin with an investigation of the constraint equations which define left-handed and right-handed chiral superfields (also known as scalar superfields). Then vector superfields are defined by an appropriate constraint equation, and the supersymmetric generalization of the Abelian gauge transformation is discussed. Finally, left-handed and right-handed spinor superfields are discussed which represent the components of the su- persymmetric field strength for an arbitrary vector superfield. Chapter 8 deals with the construction of supersymmetri grals. We begin with the definition of integration over Grassmann numbers. Then Lagrangians are constructed from scalar superfields and from vector superfields (i.e. the supersymmetric field strength). The case of the for- mer is shown to contain the Wess-Zumino model as a special case, whereas the case of the latter yields the supersymmetric generalization of the pure Maxwell theory (i.e. with no interaction with matter fields) which contains in addition to the massless vector field also the massless spinor field of the photino. Chapter 9 deals with the spontaneous breaking of supersymmetry. For the convenience of discussions the concept of superpotential is introduced. In view of the necessity of evaluating action integrals over superspace an equiv- alent and convenient Grassmann projection technique is developed. Some general aspects of spontaneous symmetry breaking are then discussed and, in particular, the Goldstone theorem is established for the general case of the breaking of supersymmetry and some other symmetry. Finally, the O’Raifeartaigh model, which is a specific theory involving three scalar super- fields, is considered and the spectrum resulting from the spontaneous break- ing of supersymmetry is investigated. In this case supersymmetry breaking results from the nonvanishing vacuum expectation value of some auxiliary field of a superfield. Finally, in Chapter 10, we consider supersymmetric gauge theories. In- troducing first global and local U(1) gauge transformations of scalar super- fields and the corresponding supersymmetric version of minimal coupling, action inte-Introduction 5 we consider super quantum electrodynamics. We then investigate the Fayet— Iliopoulos mechanism of spontaneous breaking of supersymmetry in which the latter results from the nonvanishing vacuum expectation value of the highest order component of a vector superfield. The last section contains a brief introduction to non-Abelian gauge transformations for superfields with the appropriate tensorial transformation properties.Chapter 1 Lorentz and Poincaré Group, SL(2,C), Dirac and Majorana Spinors 1.1. The Lorentz Group A point in the spacetime manifold! is denoted by (2) = (0°, 21,27, 2°), where «° = t, and z!,z?,z3 are the space components of the four-vector x!. This space is called Minkowski space and denoted by M4. The laws of physics, formulated in Mq, are invariant under the Lorentz group. Transformations of this group are linear transformations acting on four-vectors” cl! = AM oY, (1.1) leaving the quadratic form v= ah ay = Qvale” = (2°)? — (x)* (1.2) one ‘ i: ! oP = gal, = myo!” = ny AM ye?” 2" = Tyran". ‘Secs. 1.1 and 1.2 serve mainly the purpose of completeness, to define notation and to recollect some formulas which will be needed later in the text. The reader familiar with Secs. 1.1 and 1.2 could start immediately with Sec. 1.3. Readers unfamiliar with the subjects treated here can consult [75], Chap. 12, or {98] for more details. *In this book we adopt the Einstein summation convention 78 CHAPTER 1 Lorentz and Poincaré Group, SL(2,€) and Spinors Hence the Lorentz transformations A“, have to satisfy the condition: Nw A" A”, = Npr- (1.3) Here eae Oe. OF 00 (m=1) 9 <1 6 (1) O00 1 is the metric tensor; it lowers indices and its inverse n“” raises indices. Proposition 1.1: The constraints det A= +1, JA%| 21 (1.5) define four disconnected pieces in the parameter space. Proof: The determinant of a product of matrices is the product of the determinants. Hence, taking the determinant of Eq. (1.3) yields (det A)?=1 or det A= +1. Taking the 00-component of Eq. (1.3) we obtain: noo = Mav AMoA” = nooA°oA%s + nis‘ A's = (A%)? a: (A%)? or . 5 (A%)° = 14 (A*)°. Hence (A°%)? > 1. The second of constraints (1.5) distinguishes so-called orthochronous Lorentz transformations with A°y > 1 from non-orthochronous Lorentz transforma- tions with A°) < -1. Proposition 1.2: The matrices (A“,) form a non-compact Lie group, the Lorentz group? L = O(1,3;R) = {A € GL(4,R) | AA = 9} Recall the notation for the groups means the following: The O stands for orthogonal which is related to the fact that the transformations that it represents preserve the orthog- onality of coordinate axes. The numbers 1,3 stand for the fact, that the transformations obey Eq. (1.3) with the metric +1 for the time component and a ~1 for each space com- ponent. Later we shall encounter special orthogonal groups denoted by $O(1,3;R). Special means determinant equal to +1, and the geometric meaning is that orientation reversing transformations such as reflections are excluded.1.1 The Lorentz Group 9 with Lie algebra o(1, 3; R) = {8 € Maxa(R) | AT = —nAn}, where GL(4;R) denotes the set of all invertible 4 x 4-matrices with real entries,4 and M4,4(R) is the set of all 4 x 4-matrices with real elements. Proof: From Lie algebra theory we know that each A € O(1,3;R) can be written in the form® A(t) = exp(t2), where ¢ is a real parameter and & € o(1,3;R) is an element of the Lie algebra. Matrices of O(1,3;R) are subject to the condition AT(t)nA(t) = Inserting the above expression we obtain: [exp(¢%)] "n[exp(ea)] =n, and considering d ale 5 { lexn(e29] nfexp(t%)] i, =0, we obtain the conditions for the Lie algebra elements, since the Lie algebra of any Lie group is isomorphic to the tangent space at the identity of the group. It follows that 0= 4 fexp(+2)] " nexp(t2) + [exp(ea)] " ne exp! cea] , t=0 which leads to AT + 1A =0, or AT = —HAn, VAE o(1,3,R). In summary we have the following classification: Let A be any invertible 4x 4-matrix with real elements, i.e. A € GL(4,R), then: “The set GL(4;R) forms a group and is also called the general linear group. °See for example W. Miller [72], Chap. 5 or V.S. Varadarajan [115], Chap. 2.10 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors (i) The full Lorentz group is L:= O(1,3;R) = {A € GL(4,R) | ATA =n}. (ii) Proper Lorentz transformations are Ls = SO(1,3;R) = {A € O(1,3;R) | det A= +1}, Ls being a subgroup of L. (iii) Improper Lorentz transformations are L_ := {A € O(1,3;R) | det A= —1}. L_. is not a subgroup of L, since the identity element is not an element of L_. Note, however, that discrete transformations such as time- or space-reflections are elements of L_. (iv) Orthochronous Lorentz transformations are Lt = {A € O(1,3;R) | A°y > +1}. L’ is a subgroup of L. (v) Non-orthochronous Lorentz transformations are Lt = {A € O(1,3;R) | A°y < -1}. (vi) The restricted Lorentz group is LY, =Ltn ky = {A € O(1,3;R) | det A = +1, A% > +1}. This subgroup of L is also called the proper orthochronous Lorentz group; it does not contain time- or space-reflections. Remark: Lorentz transformations which are orthochronous map the for- ward light-cone onto itself, and A € L‘ maps the backward light-cone onto itself. Generators of the Lorentz group In the neighborhood of the identity 1s0(1,3;n), 2 Lorentz transformation A € L*, can be written in the form A=agxg tu. (1.6)1.1 The Lorentz Group u where 14.4 denotes the 4 x 4-unit matrix and w is a 4 x 4-matrix with infinitesimal parameters. The aim is to obtain the constraints which these parameters have to satisfy in order to obey Eq. (1.3). Inserting expression (1.6) into Eq. (1.3) we obtain MwA? Ag = tw {5"p + wt} {64 + wy} = v8 Og + Mw Se pws + Mv dg : = Noo + Wpa + Wop = No: Hence: pe = —Wops (1.7) i.e. the matrix w with its infinitesimal parameters is antisymmetric in the indices p and o. Now at 6 ge OD) Ata” us) (+w)4ja” = ch +uh a”, and the variation of the four-vector z* due to infinitesimal transformations is: bat i= o't — gt = ot oY with antisymmetric infinitesimal parameters as demonstrated in Eq. (1.7). On the other hand we may consider the vector representation of the restricted Lorentz group L}, and we write A € L', in the form v My= [exp(— 0? Mpa)", (1.8) where the 4x 4-matrices (Mp) constitute a basis of the Lie algebra o(1, 3; R), to be verified later. (Mo) are antisymmetric in p and o and the factor i is chosen in such a way that the (M,,) are Hermitian. For infinitesimal transformations, i.e. w°’ infinitesimal, we consider Sah = g't — oh = MMe’ — oh as) [exp(— 510” Myo))" 2” — alt = {axa - 507 Mo Ys" -t= 50" (Mp) 0", (1.9) and we conclude : wt, = 50 (Moo) (1.10) Therefore the generators Mp, of the Lorentz group have the matrix form: (Mpo)"y = tnov')" — novo."). (1-11)12 CHAPTER 1 Lorentz and Poincaré Group, $L(2,€) and Spinors Checking: ‘ L ae (1.7) yuh? (Myo), = 54"? (nov, — Novba!) = s(wh, — wy!) =" why. We now derive the matrix form of the generators of the Lorentz group, i.e. Eq. (1.11), explicitly. According to the definition of the Lie algebra o(1,3;R), any 2 € o(1,3;R) satisfies the relation AT = —y%An. Explicitly this condition implies for the elements of such a matrix: 7 400 Ao1 402 493 @10 G11 G12 413 429 21 422 a3 430-431 432 433, 1 ‘90 401 02 ao3\ /1 oe -1 0 410 441 412 443 -1 0 0 -1 429 21 22 a3} | 0 -1 —1/ \a30 431 432 433 a which leads to the set of equations: 499 = 11 = 22 = a33 = 0, 219 = 401, 420 = 492, 430 = 403, @21 = —@12, 431 = —@13, @32 = — 493. Therefore a typical matrix % € 0(1,3;R) has the form: 0 aor doz 03 a 0 =ai2 a3 a2 a2 0 agg 493 413 230 We choose a basis of the Lie algebra o(1,3;R) of the following form: 000 0 0 0 0-0 000 0 0001 lo og 1 feealo 60 oe 0.0 17 0 0-1001.1 The Lorentz Group 13 0 0r807.0 0100 00-10 1000 Ms=1o 1 0 of “10 0 0 of? 00 0 0, 0000 0010 en eeenacee 0000 0000 eo ool gg Go 0000 170.-0..0. The three matrices Mj,i = 1,2,3 generate the rotation group SO(3;R) as a subgroup of SO(1,3;R), the three matrices Nj,i = 1,2,3 generate Lorentz boosts. As can be shown by explicit calculation, these generators obey the following commutation relations: (Mi, Mj] = €ijxMe, [Nis Mg] = —eijeNes INi,Nj]=—eijeMay 6,58 =1,2,3, where cise is totally antisymmetric in i, j,k. The matrices Mj,i = 1,2,3, are antisymmetric, 7.e. Mj! = -Mi, whereas the Lorentz boost generators Nj are symmetric: NJ =N;. i We now construct Hermitian matrices J=iM, 1=1,2,3, with J] = (iM)! = iM! = iM = J, and anti-Hermitian matrices Ki:=iN, 1=1,2,3, such that K} = (iN)t = -iN] = -iN, = —Ki. These matrices obey the following commutation relations: (Jis Ie] = teenie, [is Kr [Ki, Kj] = —teijede. = teimKm,14 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors The first commutation relation shows that Jj,i = 1,2,3, generate the ro- tation subgroup of L1. Usually physicists take the matrices Kj and Jj; as generators of Lorentz boosts and rotations respectively. Starting with matrices K; and J; one can construct a covariant formalism, defining an antisymmetric 4 x 4-matrix M,,, 12, v = 0,1, 2.3: €4ijMij = Jk, (6, j,.# = 1,2,3 in cyclic order), Mo :=—Ki, (i= 1,2,3). Explicitly: 0 -K, -Ky -K3 oh 0e el 0, Kk3 Jp -I 0 (1.12) Proposition 1.3: In covariant notation the matrix M,, is given by Eq, (1.11), ie. (Moo), = i(nov5p" — Nov5q")- Proof: In order to prove this we consider separately: (i) p=0,0 =1. Using Eq. (1.11) we have (Moi) = imo" ~ nov5,"), and the nonvanishing elements of the matrix Mo; are (Mo1)°; = -i, (Mo1)"y = -i. On the other hand using Eq. (1.12), (Moi)", = —(Ki)*, = -i(M)* and the explicit form of N, gives the only nonvanishing elements (M1)°, = 1 = (Mi)'o- Hence (Mo1)°, = —i = (Mo1)'o-1.1 The Lorentz Group 15 (ii) p= 2,0 =3. Using Eq. (1.11) we have (Ma3)", = i(nsv5" — nav53"), and the nonvanishing elements of M23 are (Mz3)?5 = i, (Mz3)°p = +4. On the other hand using Eq. (1.12), (Mos), = (AE, = ta)", and from the explicit form of M, we obtain (Mzs)*s = i(Mi)’s = -i, (Mz3)*x = i(Mi)*. = +4. The other matrices can be checked analogously. We now define (Myo) pv = Nur( Moo)” y = itr (Nov p” — Novda") = t(NouNov ~ NpvNon)s (1.13) where we used Eq. (1.11). Proposition 1.4: The matrices (Moo) ww] defined by Eq. (1.13), are Hermitian matrices, i.e. (Moo) wl" = (Moo) uv]. Proof: Consider [(Mpo) w)t = fi (Noutev — nw'ton) | = ~i(nounov — nvtiou) |» We observe that transposition means interchanging the matrix indices jz and vy, ive, (Moo) w]t = -i(npunon ~ Noutov) = *(Npptev ~ Novtion) = [(Mpo) ww] (using Eq. (1.13)).16 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors The Hermiticity of the matrices [(Mp¢),.] implies for the generators K; and Ji: (aw) = (uw), [Kw] = [Kye]. Explicititly we have: (i) = Mup( A)? (Ji)as = nep(Ji)s = ma2(Ji)?s = -i(Mi)s = 4, (Ji)32 = nap(Ji)’s = n3a(Ji)%2 = -#(Mi)%) = — i. All other elements of this matrix are zero. Hence 00 0 _{00 00 (el=]o 0 0 5 00 -i 0 Similarly one shows that 000 0 0 0 00 000 -i 0040 (wl=]o 00 of Mswl=lo ¢ 0 9 Ot 02.0) 0000 For the boost generators we have (Ka) uw = Mup( Ki)’, (Ki)or = nop(i)": = no0(F1)°, = 1M) = 4, (Ki)io = mp(K1)%o = mu (Ki)'p = —i(Mi)' = -4, and all other elements of this matrix are zero. Therefore the explicit form of (Aa) po) is: 0 100 - 000 (Kw) = O00 (eee Vee a) The boost generators Kp and Ky are given by 0 0% 0 0008 0000 0000 [(Ko)w] = |; 0 0 of? [Ks)uu] = O00 6 O07 070 1 0001.1 The Lorentz Group 17 Hence in this form the six matrices {Khae and (hire form a set of Hermitian generators of the Lorentz group. The Lie algebra o(1,3;R) describes the Lorentz group locally and is determined by the com- mutator of the basis elements. Proposition 1.5: The generators M,,, of the Lorentz group obey the following com- mutation relation: [Mus Moo] = ~i€tupMve — MueMup ~ MpMyo + TveMyp)- (1.14) Proof: Using Eqs. (1.11) and (1.13) we have: [Mus Mpelag = (Muv)ar(Mpa)%3 ~ (Mpo)ar(Myw)"y 4 (Nuala ~ Murtiva) i (5p "Nop ~ Mp6") ~i(Npattoy ~ Novtioa)i (Sve ~ MwaSv”) = =(Matrptep ~ NueNveNps ~ Nuptvatop + Nuotwvatea) +(NpatouMB ~ NpaNovNup — NouNeate + NovNoaNys) ' [ruc (types — MwoNps) ~ twa (Mptas ~ Nuotps) Mpa (Noptvs — Novlys) + Noo (pump — ora) | —inuol®) (Nvattes ~ Nvptioa) + iyo (4) (Mvatps — Nats) +invel®) (Nueces — Naas) ~ *tva(t)(Nuates — NoaNus) = ~inyp(Mve) ag + ine (Mvp) ag tinue Mpa) ag ~ itve(Mup) ag: Dropping the matrix indices a, 6, we obtain Eq. (1-14). Next we want to rederive (as a consistency check) the commutation rela- tions of the generators A; and Jj, starting from the commutation relations Eq, (1.14). We had: Mn = €mnidi, (1.15) Moi = -Ki. (1.16) Equation (1.14) reads for the indices p = p = 0, =i,0 = 9; i,j =1,2,3: 1,16) i [Mois Mo; OD) KG, Kj] = —i(nooMij — 103Mio — nioMoj + 1:7 Moo) 1.14) iz ia C29 _ in Miy = ~ieign Tee18 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Hence: [Ki, Kj] = tein Je. (1.17) For p = 0,v =i,p=k,o =1; i,k, = 1,2,3 we have: 115 Fi (Moi, Mul’ [Kis €ktmdm] = —i(no Mit ~ nor Mix — nix Mor + nuMox) ou o =" ~inkKi + mK, = din — Fuk), where in the last step we used nig = ~Sjx. Hence: exim|Kiy Im] = —i (Sik Ki — 6 Ke). (1.18) Then using €kin€kim = 25nms (1.19) we obtain (with Eqs. (1.19) and (1.18)) 26nm{Kis Im] = 2[Kis Jn] = ~tepin(OikKi ~ bnKe) = —teiin Ki + terinKe = icin Ki. exin€kim{Ki, Im] Hence [Ki, Je] = ies K1, (1.20) and finally we have to evaluate Eq. (1.14) for the case # = i,v = j,p = ko =1; i, j,k, = 1,2,3. Using Eq. (1.14) we have [Mij, Mu] = [¢iimdns€kin Ja] = €ijménin[Jms In] = —i(ninMjr — naMjx — njeMa + njMix) = ~i(nix€jta — Nit€jka ~ Nyk€ita + Nyttika) Ja- Then with Eq. (1.19) 1 Joti FBlgigmEKIn [Jis In] = bfm5gn[Ims In] = [Jy, Jg] i = — [eis ehig (Mine jla — Muejka — Nakita + Njt€ika) Jo a =-4 { —€ij FEilgejla + €ij fEKigejka + €ig fEjlg€ila — Eig s¢hjgtiba bJn a = 5{Gibr0 ~ 9795 puesta + (Sjx5 fy ~ 9)96 pk) Eka + (Buby — Sigbrp eta + (6i85j9 ~ SigSs)¢iba} Jo i . = 5 {cara + coja t+ Cafe + ofa} Jo = iesoaes1.1 The Lorentz Group 19 where we made use of nij = —6jj and €igk€itm = Ojt5km — Ojmokt- Hence: [vis Jj] = teijnJe- (1.21) Equation (1.21) defines the rotation group $O(3;R) as a subgroup of L1, whereas Eq. (1.20) states that K is a vector under the Lorentz group. The minus-sign in Eq. (1.17) is significant; it expresses the difference between the non-compact group SO(1,3;R) and its compact form SO(4;R) or between SL(2,C) and SU(2,€) x SU(2,C), since locally homomorphic Lie groups® have homomorphic Lie algebras. In order to be able to classify the irreducible, finite-dimensional, non- unitary representations {103] of the restricted Lorentz group L1, we have to change the basis Kj, J; of the Lie algebra so(1,3;R) by introducing the complex linear combinations (1.22) (1.23) One can verify, using Eqs. (1.17), (1.20) and (1.21), that these non-Hermitian generators decouple the commutation relations of K; and J; so that (Si, 55] = tein Se, [Tis Tj] = teignTe, (1.24) (Tj, 5;] = 0. This means that the generators S; and Tj obey the commutation relations of the Lie algebra of SU(2,C). In addition, the commutation relations of Eqs. (1.24) show that the Lie algebra so(1,3;) decomposes into the direct sum of two su(2,C) Lie algebras. However, this decomposition holds only for the complexified Lie algebra so(1,3;R)©, i.e. considering the set of real 4x 4-matrices % satisfying AT = — Hn as a complex vector space, this allows complex linear combinations of the form S; and T;. Hence the decomposition $0(1,3;R)© & su(2,€) x su(2,C) ®The groups SO(1, 3; R) and SL(2, C) are locally homomorphic as we shall see explicitly in Sec. 1.3.3, as well as the groups SO(4,R) and SU(2,C) x SU(2,C).20 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors is valid only for the complexified Lie algebra of the Lorentz group. However, we can use the classification of irreducible representions of the complex Lie algebra so(1, 3; R)© to find the irreducible representations of the real Lie alge- bra so(1,3,R) © si(2,C), since there is a one-to-one correspondence between representations of a complex Lie algebra and representations of any of its real forms [19,72]. But the classification of finite-dimensional irreducible representations of the algebra su(2,C) x su(2,€) is well known. According to the theorem of Racah,’ there is one Casimir operator for every su(2,€) subalgebra, i.e. 3 3 58 and SONG i=l i=l are Casimir operators, commuting with any element of the algebra, with eigenvalues n(n+1) and m(m+1) respectively. Here n and m are eigenvalues of $3 and T respectively with values n,m = 0,1/2,1,3/2,.... Therefore we can label representations of so(1,3;R) by the pair (n,m), and since J3 = 53+ Ts, we can identify the spin of the representation with n +m. The dimension of the representation space is (2n+1)-(2m+1) for an (n, m)-representation of the Lorentz group. It is important to note that the two su(2,C) subalgebras are not inde- pendent since they can be interchanged by the operation of parity. Parity acts as follows on rotation and boost generators: Ii S, Ki + -Ki, and Eqs. (1.22) and (1.23) show that parity transforms S; into T; and 7; into S;. In addition the operation of Hermitian conjugation also interchanges S; and T;, since, as demonstrated above, J; and K; can be chosen to be Hermitian: 1 A 1 i S] = 5h + 1K)! = 5(Si- 1K) = Ti, 1 1 : qT} = Pica -ik,)t = g(Ji + 1K) = Si. Hence, the parity operation is equivalent to Hermitian conjugation. As examples we consider the following representations: TSee e.g. W. Miller [72], p. 3951.2 The Poincaré Group 21 (a) (0,0) with total spin zero is the scalar representation; the dimension of the representation space is one. (b) (1/2,0) with total spin 1/2 is called the left-handed spinor representa- tion; the dimension of the representation space is two. (c) (0,1/2) with total spin 1/2 is called the right-handed spinor represen- tation; the dimension of the representation space is again two. The handedness is a convention. In subsequent sections we shall discuss these two spinor representations in great detail. Since parity switches 5; to T; and T; to Sj, representations of the Lorentz group in general are not parity eigen- states. In particular, the left-handed spinor representation (1/2,0) trans- forms under parity into the (0, 1/2) representation and vice versa. Therefore to obtain a representation such that parity acts as a linear transformation, one has to consider the direct sum of the spinor representations (b) and (c), (1/2,0) ® (0, 1/2), which yields a Dirac spinor representation. This repre- sentation of the Lorentz group will be considered in detail in Sec. 1.4. The importance of the representations (b) and (c) is due to the fact, that any other representation of the Lorentz group can be generated from these spinor representations. In Sec. 1.3 we shall consider a few examples explicitly. For instance, the Kronecker product of the representations (b) and (c) (1/2, 0) ® (0,1/2) = (1/2, 1/2) gives a spin 1 representation (see Eq. (1.152a)) with four components, and the Kronecker product of two left-handed spinor representations decomposes into a scalar representation and a spin 1 representation, given by an anti- symmetric, selfdual second rank tensor (see Eq. (1.152b)), i.e. (1/2, 0) @ (1/2, 0) = (0,0) ® (1,0). 1.2. The Poincaré Group As stated above, the Lorentz group leaves the interval (z — y)® in Minkowski space My invariant. On the other hand, the translations ch —> gt = ot + at, where a! is a constant four-vector, also leaves the length squared (x ~ y)? invariant. ‘This leads to the definition of the Poincaré group P as the group22 CHAPTER 1 Lorentz and Poincaré Group, $L(2,C) and Spinors of all real transformations in Minkowski space of the form:® ot > ol = Myo” 4 al, (1.25) which leaves the length squared (a — y)? invariant. Definition (1.25) leads to the following composition law for the elements of P: Let ] Aye+ai, al" = Aga! + ag = Ag{Air+ai} +a == AgAja + Aga; + ap. Writing (A,a) for an element of P, we have for the composition of two Poincaré transformations: (Ag, a2) 0 (Ai,a1) = (AzA1, Azar + a2). (1.26) This demonstrates that the Poincaré group P is the semi-direct product 3 L @ T4 of the Lorentz group L and the translation group in four spacetime dimensions T, in Minkowski space. Similarly, as for the Lorentz group, the Poincaré group P decomposes into four pieces identified by det A and A°; ie. Pi, Pl, pt, pt ‘The identity element of the Poincaré group P is the element (114,4,0), and the inverse of the transformation (A,a) € P is (A~!, —A~1a) such that (applying the composition (1.26)) (A,a) 0 (A7*, Aa) = (AA7?, ~AA“*a +) = (Laxa,0), (A7}, —A71a) 0 (Aja) = (AT1A, ATta — Awa) = (Lax, 0). The Lie algebra of P! is determined by the commutation relation (1.14) of the Lorentz group ce the trivial commutation relation of the translation group (observe that the translation group T, in Minkowski space is Abelian), and the commutator of translations and Lorentz transformations, stil! to be determined. In order to obtain this commutator, we consider a faithful representation of P}: (A,a) — g(A,a) (1.27) °In the mathematical literature transformations acting on a manifold that preserve the distance between pairs of ponts are called isometries. Therefore, the Poincaré group is the isometry group of Minkowski space, see e.g. (72], p. 20.1.2 The Poincaré Group 23 in a vector space V such that 9(Ag,a2)9(A1,@1) = g(AgA1, Aga; + a2), (1.28) g" (Aa) = g(A7!, -A~1a). (1.29) This means, g is a bijective linear map of a vector space V onto V, satisfying Eqs. (1.28) and (1.29). Infinitesimally we can write? Ard) = By ~ Supe M™ + ia, PH, (1.30) where wpg = —Wgp are six infinitesimal parameters leading to an infinitesimal Lorentz transformation A € L{, and a, denotes four infinitesimal parame- ters, leading to an infinitesimal translation. Now M?? = —M°? and P¥ are generators of Lorentz transformations and translations respectively in the corresponding representation. Consider!® g71(A,0)g(A',a')g(A,0) °2 97 (A,0)g(A'A, a!) g(A~!,0)g(A'A, a!) 29) g(a 1ATA, A“*a'). (1.31) (1.29) For infinitesimal (A‘,a’) € Pt the left hand side of Eq. (1.31) gives (1,30) 1A, 0)g( A’, a")g(A, 0) go (A, ftv 5 — Sul,MM + ial, Pith (A, 0) Pet = ty ~ ufo (A, 0)" 9(A, 0) +iaj,g~'(A,0)P#9(A, 0). The right hand side of Bq. (1.31) may be expanded as g(ATIA'A, Anta’) = ty - 5(A TW Apo MA? +i(A71a'))P? 24) 50M) fey ANY Mor +i(A!) fal, PP =1y- 5 pAM AY, M? + ia!,A", PP. ®°We denote by lv the identity map of the vector space V. 20We apply the properties (1.28) and (1.29) of the faithful representation g of the Poincaré group.24 CHAPTER 1 Lorentz and Poincaré Group, $L(2,€) and Spinors Hence we obtain: g7(A,0)M#Yg(A,0) = AMpAY, MP7, 7 M(A,0)PAg(A,0) = AMP. 1 The first equation states that M“” is an antisymmetric tensor operator under L1; the second relation shows that P! is a vector operator under L1. Now consider infinitesimal A € wt; for Eq. (1.32) we obtain: go \(A,0)M#"9(A, 0) = g(A7!, 0) M*¥g(A, 0) = (tv+ apo”) mM (ty - Supe”) = MM 4 50 [mer me), On the other hand, the right hand side of Eq. (1.32) can be expanded as: MAY, MP? it (5%, + wh) (8%, + wg) MP = MY + wh oY MP + iw" MP = MY +uH,M™ +" ,MM = MM + 9M gp + 14! Wyg MY? = MMW 4y Slot ope + 1f"?tipg MY t1Pwpe MY + 1°? wopM*?} = Me + Feo — nf? MY =n’? MHP + nM}. Hence [MP MH) = E(t MY — MO — 9 MH + MH), or (Mm, MP7] = I(t? MY? — nH? MY? — 9” MH? +7 MHP), In this way we have rederived the commutation relation of the Lie algebra s0(1,3;R), ie. Eq. (1-14).1.2 The Poincaré Group 25 The second of Eqs. (1.32) leads to i : = Pe dele, PA 97 "(A,0)PP9(A,0) = (tv + SuuM) PP (ty — SeeM) +A?,P” — (using Eq. (1.32)) = (0, 408) PY = PP + wp” = PP + ayy PY 1 = PP+ gee PY +P wrpP"} 1 = PP+ etn {nf PY = 9 PHY, Therefore we obtain for the commutator of the operator My, with P,: [Mt P?] = —i(nt? PY — n°? PH). (1.33) In summary we see that the Poincaré algebra is given by the following set of commutators: (1.34) Proposition 1.6: The square of the energy-momentum vector P? = P,P! is a Casimir operator of the Poincaré algebra (1.34), i.e. [M,v, P?] = 0, (1.35) [Pu; P?] =0. (1.36) Proof: The verification of relation (1.36) is trivial. Hence we consider the26 CHAPTER 1 Lorentz and Poincaré Group, $L(2,C) and Spinors relation (1.35). Using Eq. (1.33) we have: [My P?] = [My P?Po] = [Myo P?] Pp + P? (My Po] = (My, n?? Po] Py + P? [Myr Po] i uo Pv — va Pu) Pp — P?i(MupPv — wp Pu) inf Typo PLP 5 + iP? Ma PuPy — iPP yp Ps + 1P ompR = ~i6?,, P,Pp + i? PyPp — iPyPy + iP, Py = -i[P,, P,] -i[P,,P,] =0. The second Casimir operator of the Poincaré algebra (1.34) can be con- structed from the Pauli-Ljubanski polarization vector!! defined by Wy €uvpa PYM", 1.37) ’ Hu p 2 where €ypo is the Levi-Civita tensor with eo123 = +1. Proposition 1.7: The Pauli-Ljubanski vector is invariant under translations, 7.e. [Pu Wi] = 0. (1.38a) It can be written in the form W* = [I, P*], (1.38) where i Ime » yer €uvpoM*” MP, and is a vector under L1: [Mw Wp] = —i(MmpWr ~ tupWy)- (1.39) Proof: We first demonstrate Eq. (1.38a). With the help of Eqs. (1.37) and ‘gee the discussion in R. Penrose (87], Chap. 22.121.2 The Poincaré Group 27 (1.33) we have: az Hi [Pas Wo] = 5 [Pas vpor PPM] = Seupar [Pun PPM] = Fevportin{ [P, PPM + PPP, Mer] oa 5vperthen PP {PT et ney i at 3g leva PPPT i evpauP?P?] . 5 [evnur PPP + €urupP™P? = evpoy PPP? — evopP? P?] =0. Proof of Eq. (1.38b): [P) = [fens sM™M, PH = 5 coors [e8 M5, PX) - Faas {Mee (MS, PH) + (Me, PH) } = jase {M(t PF — PY) =i(n Pe — nha) ms a {093517 MPP — copys! MP? Hel ggg PPM ~ esl gPOM™} = je age MOP + Zeta PAM. Now owing to the commutator given in Eq. (1.33) we can write: Fl eae ie ae and contracting with e“.95 yields 1 1 1 qe opeM PPE = Fe yggP9 MO = FelsggP°M?, so that 1 (1, P4] = ge ash OM™ =we.28 CHAPTER 1 Lorentz and Poincaré Group, S1L(2,€) and Spinors In order to prove Eq. (1.39) we consider the following commutator and use the Jacobi identity: [mew] = [ae 7, P| = -[r, [P?, me i [P*, [m7] = —[L,i(nt? P’ - 1’? P*)| -0 = ~ a(n’? [T, PY] - 9? [T, P¥]) = ~i(n"?W" — PW"), where [M“’, 1] =0, since J is invariant under Lorentz transformations. Proposition 1.8: The square of the Pauli-Ljubanski vector is given by the expres- sion W?=W,W" = ~ 5 MM PP + MP? My P,P’. (1.40) Proof: W? = WW = WW 1 1 = af" (Senapy POMP”) Sevpor PPM} 1 OT = 7" a8 y€ve07 P* MP" PPM : Applying the following contraction of two totally antisymmetric tensors in four-dimensional Minkowski space, yaq€uoot = Nap (NBayr ~ NBrNye) Nao (NBpNyr — N6rte) +Nar (ngpttra ~ NBoNyp) » we find : 1 W? = ~F{nap(nsotiyr ~ Nartina) —Maa (nsethyr ~ "6-Mhrp) thar (naptne ~ Na0thr9) }P2M®? PPM" = —1{P)MorP?M® ~ PpMro PPM ~ PeMprP°M™™ +P, MzpP°M + P,Mpy PPM" — P;MopP? M7}.1.2 The Poincaré Group 29 Now using Eq. (1.34), PpMorP?M = {MorPp ~ i(nrpPa ~ NopPr)} P°M™ = Mgr P,P?M°" —i(P,P, — PrP,)M” = Mg,P?M —i[P,, Pr] M°" = M,,M P*, where we applied Eqs. (1.34) and (1.35). Using the antisymmetry of the generators M°? we get: = Py Mpr PPM?" + Po MrpP?M®" + PrMygP?M°" — P,MopP?M = -2PyMyrP?M + 2P,Myo PM = —4P,MprP?M°" = —4{ MprPo — i(troPp — Np Pr) }P? M7" = —4Mp,P,P?M°" + 4iP?M," — 4iP, PM?" = —4MprP, PM’, since M,” = 0, and because the antisymmetry of M7" implies that P;P,M*° vanishes. In the same way one can show that ~My PP? M7" = —4My,M P, PP, and therefore Ww? =W,W4 = = 5 MMP? + M"’My P,P’. The importance of the Pauli-Ljubanski vector is due to the fact that its square W? is the second Casimir operator of the Poincaré algebra. Proposition 1.9: ‘The square of the Pauli-Ljubanski vector, W?, commutes with the generators of the Poincaré group P, and Myy, i.e. [Mu,W?] =0, [P,,W?] =0. (1.41) Proof: Using Eq. (1.39) we have: [My W?] ad [My WoW?] = [Muv,W,]W? + W? [Myv, Wo] = ~i(MpWo — typWyu)W? — iW? (pW — tpWu) = -i(WLW, — WW, + W.W, - WW) =0.30 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors With Eq. (1.38a) the second of relations (1.41) is trivial, ie. [Pu W?] = [Pu W,]W? + W?[P,, Wp] = 0. Proposition 1.10: ‘The Pauli-Ljubanski polarization vector has the additional prop- erty W,P# =0. (1.42) Proof: (1:37) A uope PY MP PH ? : ; 5fnvee PYPHM™ + Seva PY PP 7 5 twp PY PtP? = ewe PYPHM + Févugo PMPY MP? W,,P# (1,34) 4 +5 euap! POPP — Sey hg PY PP ww! a (1.34) = Foye [PY Pr] Me 2 9 The representation theory of the Poincaré group has been discussed ex- tensively in the literature using the formalism of induced representations and the concept of little groups.!? We do not go into details here, therefore, and simply quote the following results [103], [106]: The unitary (infinite-dimensional) representations of the Poincaré group can be split into three main classes. These are: (a) P?= P,P! =m? >0; W? =—m?s(s +1). (1.43) The eigenvalue of the second Casimir operator W? is —m?s(s + 1), where s denotes the spin which assumes discrete values 3 = 0,1/2,1,3/2,.... In the mathematical context, little groups are called isotropy groups or stability sub- groups. See e.g. the books by V.S. Varadarajan [115], Chap. 2.9, T. Brécker and T. tom Dieck (19], and A.O Barut and R. Raczka [8], Chap. 16.1.2 The Poincaré Group 31 From Eq. (1.42) one deduces that in the rest frame (P# = (m,0)) the zero component of the Pauli-Ljubanski vector must vanish, and the space components in the rest frame are given by 1 Wi= einge POS, such that W? = —W? = -m’s?, (1.44) where 1 siz 3S (1.45) is the spin operator. This representation is specified in terms of the mass m and spin s. Physically a state in a representation (m, s) corre- sponds to a particle of rest mass m and spin s; moreover, since the spin projection s3 can take on any value from —s to +s, massive particles fall into (2s + 1)-dimensional multiplets. P?=0; W?=0. (1.46) In this case, W and P are linearly dependent: Wy = APy. (1.47) The constant of proportionality is called the helicity and is equal to +s where s = 0,1/2,1,... is the spin of the representation. The time component of W# is we = SOP Msp G2) py, (1.48) so that Eq. (1.47) implies P-J A= (1.49) which is the definition of the helicity of a massless particle. Examples of particles which fall into this category are the photon with spin 1 and helicity states +1, and the neutrino with spin 1/2 and helicity states £1/2. (1.50)32 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors This type of representation describes a particle of rest mass zero with an infinite number of polarization states labeled by the continuous variable p. These representations do not seem to be realized in nature. Remark: For this case the calculations of case (a) do not work since we cannot make a transformation to the rest frame. However, we can always transform to a system where Py = (Po, 0,0, Po). If w, is the eigenvalue of the Pauli-Ljubanski vector W,,, Eq. (1.42) implies 0 = wy, P# = WP? — wi P?, ie. we au w and wy 2 fcr 2 ce 2 PY ayldy = Wy — w? = — (uit + wR) = ~p?. Thus in this case the eigenvalues of the Casimir operator W? can as- sume any value. 1.3 SL(2,C), Dotted and Undotted Indices 1.3.1 Spinor Algebra We consider the special linear group in two dimensions with complex param- eters, SL(2,C) = {M € GL(2,C) | det M = +1}. (1.51) A linear representation of this group is a map from the group $(2,€) into the automorphim group of a certain vector space!’ F. This means M €SL(2,C) —+ D(M). (1.52) The automorphism group being defined as the set of linear bijective maps from F to F, the group multiplication being the composition of maps. As “The reader who wants to refresh his memory of definitions of mathematical terms without delving into mathematical texts is advised to consult the article by A.S. Sciarrino and P. Sorba [104] which is a “Junior Dictionary” of group theory concepts commonly met in particle physics. A rigorous mathematical treatment of these subjects can be found in the textbook by S. Lang [66]1.3 SL(2,C), Dotted and Undotted Indices 33 usual we demand the representation properties D(tsz(2,¢)) = Ir, (1.53) D(M1)D(Mz) = D(M, + M2), YM1, M2 € SL(2,€), (1.54) where 1572,¢) is the unit element of the group SL(2,C) and 1p is the identity map in the vector space F’, Let be any element of F and {@;}i<1,..,aim F the canonical basis in F. Then dim F b= YS dnbrs (1.55) n=l and representation matrices act on in the following way: dim F D(M)v= D> vnens (1.56) n=l where dim F $= YS D,(M)wi. (1.57) i=l The set (D,,"(M)) are dim Fx dim F-dimensional matrices called repre- sentation matrices; dim F is called the dimension of the representation. Two representations D“, D®) are called equivalent, if an invertible dim Fx dim F-matrix U can be found, such that D'Y(M) =UD®)(M)U-!, U €GL(F,K), K=Ror. (1.58) The special linear group SL(2,C) admits two inequivalent spinor represen- tations, as we shall see:!4 (a) The self-representation: The self-representation is defined by D(M):= M, VM € SL(2,C). (1.59) The dimension of the self-representation is therefore two. According to Eq. (1.57) elements of the representation space € F transform under the self-representation as!® Vy = My? bp, A,B=1,2. (1.60) Ty the context of supergravity and superstrings one considers supersymmetric theories in more than four spacetime dimensions. For a discussion of properties of spinors and y-matrices in an arbitrary number of spacetime dimensions see A. Pais [85] or P. van Nieuwenhuizen (114). ‘We adopt the Einstein summation convention also for spinor indices A, B,... which means summation over repeated indices.34 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors In the literature the elements ~ € F transforming according to Eq. (1.60) under SL(2,C) are called left-handed Weyl spinors or covariant Weyl spinors, and this representation!® is denoted by (1/2, 0). (b ‘The complex conjugate self-representation: The so-called complex conjugate self-representation is defined by D(M) := M*, YM € SL(2,€), (1.61) where M* means complex conjugation. As in case (a), the dimension of the representation space is two. We call the corresponding represen- tation space F’. Elements of this representation space are denoted by @ and transform under $L(2,C) according to Bia (M1) 05 AB=1,9, (1.62) where $y € F. The elements #4 € F transforming according to Eq. (1.62) are called right-handed Weyl spinors, and the representation is denoted by (0, 1/2). Left- and right-handed Wey] spinors are related by complex conjugation, i.e. taking 4 € F, then (see also Eq. (1.200)) (pa) =P € P. (1.63) It should be observed that this equation does not exhibit the same index structure on both sides. As long as we deal only with Weyl spinors this does not give rise to difficulties. A relation consistent with Eq. (1.63) which does exhibit the same index structure on both sides will be given in Sec. 1.4.4. The self-representation and its complex conjugate representation are inequivalent, i.e. it is not possible to find a 2 x 2-matrix C, such that M=cm*c". Proposition 1.11: The representation D(M) = M7" (1.64) is equivalent to the self-representation given by Eq. (1.59). 1See e.g. the article by es [77] or the textbook by R.U Sex! and H.K. Urbantke {106], Chap. 8.3.1.3 SL(2,C), Dotted and Undotted Indices 35, Proof: We have to show the existence of a matrix € such that eMe!=M~"", — € €GL(2,€). (1.65) Let ca tani= (2&2) en €22)? and. es ‘a 1 é —€ Tae fee eakee (aca 12 et = (ean) dete (2 an )> and es i) Mz M22)* Then : mote My -Mr\ _ ( M2 —Ma det M \-Ma Mu -My2 Mn J? since M € SL(2,C), and so det M = +1. Then we have to show that the matrix equation 1 (eu ez) (Mur Miz) ( 22 -€12\ + ( Mar un) dete (err €22) \Ma1 Mao) \-en en -M2 Mu leads to a consistent set of matrix elements €4g. Evaluating the product of the three matrices we have: -1f@ 6\ 1 ( Me -Ma cao (2S) ( Me i). with a = Myye11€22 ~ Myz€11€21 + Maiei2€22 — M22€12€21, 2 b= —Mierrer2 + Mireri€22 — Marejg + Mageire22, 2 2 © = Mi1621€22 — Mize, + Mai €92 — Mrze22€21, 2 d = ~My €12€21 + Mizeoi€22 — Mayeizess + Mage59. This leads to the following set of equations: Mirersege ~ Mizeiie2i + Mai€i2€22 — Ma2€12€21 = Mo dete, 2 2 Myye21€22 ~ Mizea, + M2ie99 — Maze22€21 = — Miz dete, 2 —Mirerier2 + Miveriese — Moiejz + Moze12€22 = ~Moy dete, 2 —Myireizé21 + Mizeoi€22 — Morei2€22 + Maze92 = Mii dete.36 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors From the first equation we deduce by equating coefficients of Maz on both sides e=e@2=0 and — ener =dete. From the second: &, = dete, and from the third ely = dete. Then ~e12€21 = dete = ef, ke. sear nea We choose €;2 = —1 so that €21; = +1, and we have: 0 -1 = taad=() g)=er, (1.6) where the right hand side defines a matrix with upper indices. Then the matrix € with upper indices is: (<4) = e 6) =(eaa)", (1.67) such that (48) (AB)"! See! = Boxe GG )-G 3) eae =64q, cape? = 64°, ehge®> = 5°. (1.68) implies or in mixed form: With this index convention, i.e. writing ¢ with upper and e~! with lower indices, so that € plays the role of a metric, we can write Eq. (1.65) AB M,C eon = (M7")4p. (1.69) Thus we have constructed a 2 x 2-matrix ¢, such that eMe} = MT” Hence, the two representations M and M~'" are equivalent.1.3 SL(2,C), Dotted and Undotted Indices 37 Multiplying Eq. (1.69) by eg from the left and ¢2¥ from the right, we obtain: ena’? MpCecne® = epa(M'")4 pe, Sp? Mp So" = ena(M")A pe”, Mp” = epa(M~'")Ape?F, (1.70) where we used Eq. (1.68). Now, taking a %,4 € F, we know from Eq. (1.60) that this spinor trans- forms as 1.70) ine Ua = Ma? bp OD eac(MT) pe Mabe. Multiplying this equation by e”4 from the left and again using Eq. (1.68) we arrive at: ePAyl, = (MVP pe? Pb. (1.71) We next define a left-handed Wey! spinor with contravariant spinor index by YA = AB yp. (1.72) Then Eq. (1.71) reads: oA = (MTA By. (1.73) Proposition 1,12: ‘The representation D(M) =: M*T (1.74) is equivalent to the complex conjugate representation given by Eq. (1.61). Proof: As in Eq. (1.65) we have to search for a 2 x 2-matrix € such that (1.75) In the same manner as for Eq. (1.65) one can show that a c ) =: (48) (1.76a)38 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors and -1_ (0 -1 eae 6 0 ) =: (€45)- (1.76b) In index notation Eq. (1.75) is written ABM") Ween = (mT). (1.77) Multiplying from the left by €j4 and from the right by e?’, and using exe = 644, (1.78a) eager’ = 6, ic. (1.78b) we obtain: : ee (M*) 5 = ge(MI71T)S pe? (1.79) Using Eq. (1.62) we have the transformation property of dotted Weyl spinors (right handed Wey! spinor) a. hae 2 . Ba = (M*) PBs = cgg(Mi1")e OPH. Multiplying this equation from the left by #4 we obtain (1.780) (tT) E & BAG = Pe, (Mt 1TVe, DEG, » Defining dotted spinors with contravariant indices by the relation A ips P= APG, (1.80) we conclude that dotted spinors with contravariant indices transform under M*-'™ according to: B= tye. (1.81) Summarizing we have: For any M € SL(2,C) the matrix M, its complex conjugate M*, the inverse of its transpose (M‘)~1, and the inverse of its Her- mitian conjugate (M+)~! all represent the group $L(2,C). Two-component spinors with covariant and contravariant dotted or undotted spinor indices transform under S'L(2,C) as follows: Ua = Ma? de, (1,82a) WA = (MIT) gy, (1.82b) Ba = (M1) Bp, (1.820) A (MrT) gp”, (1.82d)1.3 SL(2,€), Dotted and Undotted Indices 39 The raising and lowering of spinor indices has to be understood in the fol- lowing way: (i) Given any spinor which transforms under the special linear group SL(2,€) in the self-representation, 4, we can construct a contravari- ant spinor 4 which transforms under M=17; 4 = ABypy = —vpe?. (1.83) (ii) For a Weyl spinor which transforms under M~!", $4, we have va = cape’. (1.84) (iii) Given any Weyl spinor which transforms under $Z(2,C) according to M*, $4 BF = AbG, = —FgePA. (1.85) (iv) For a Weyl spinor in the representation (0, 1/2) which transforms under (Mt)-1; U4 = eis (1.86) The raising and lowering of spinor indices is performed with the help of the matrices ¢ and e~!, which also connect the two equivalent representations M and M~'7, Hence the matrix plays the role of a metric in the spinor space F. Explicitly, the raising and lowering of the spinor indices is given by: (i) Raising undotted indices: yA = AP dp, pl = el Bop = ede = Ye, since ¢!? = +1 (see Eq. (1.67)) and P= OP by = Oy = th (ii) Lowering undotted indices: va = ean, th = apd? = ead? = since €}2 = —1, hae.40 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors (iii) Raising dotted indices: where e!2 = +1 (see Eq. (1.76a)) =i bac sic B =28F,= 2 (iv) Lowering dotted indic 7 <8 PA = CisP since €j3 = —1,¢5j = +1. This is a consistent set of equations. We now define dual spaces!?7 F* and F’* and establish the connection be- tween y* and @. As explained above, we have four types of two-component Weyl spinors transforming according to four different representations of the group SL(2,C), where M, M~'T, and M*, M*!" describe equivalent repre- sentations. (i) Any % € F transforms under the self-representation (1.59) as Ua = Ma? dp. Such spinors are characterized by a lower spinor index. (ii) From Eq. (1.64) we know that spinors transforming according to Oi = (MTA gp? form a representation of SL(2,C) that is equivalent to the self-represen- tation. Such spinors carry upper undotted spinor indices and are ele- ments of the dual vector space F*. 1THere we adopt the mathematical standard notation for a dual space which is an asterisk. This should not be confused with complex conjugation.1.3 SL(2,C), Dotted and Undotted Indices 41 (iii) A spinor transforming according to the complex conjugate self-repre- sentation is given by #4 € F and is characterized by a lower dotted spinor index, = aa Ba = (M") Pp. (iv) Finally, an equivalent representation of SL(2,C) is given by spinors which transform according to ye = (Mt ITA Such spinors carry upper dotted spinor indices, i.e. contravaraint dot- ted indices and are elements of F*. A mathematical framework for these different types of spinors is given by the following considerations. Consider F as the vector space of two-component spinors ~,4, we may construct the dual space F** in the following way. According to linear alge- bra,!® the elements of the dual space F* are linear maps ¢ from F to a field K, which we take as the field of complex numbers C: ¢:F—->C, such that for all y € F: OH) = da EC. (1.87) Hence according to our index convention we may interpret two-component spinors with contravariant undotted indices as elements of the dual space Fy ie. paek, per. In addition we know from Eq. (1.72) how we can correlate an element of F to the corresponding element of F*. The e-matrix may be considered as a map (48) :F — F*, pa > ot = AB dp. The inverse map is, of course, given by the inverse matrix, i.e. the e-matrix with covariant spinor indices: (cag) : F* —> F, ot > ba =eany?. '®See the book by S. Lang {64] for a mathematical discussion of these aspects.42 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Interpreting the composition on the right hand side of Eq. (1.87) as matrix multiplication, spinors with upper undotted indices represent rows and those with lower indices represent columns. In the same way one can consider the space F* as the vector space of two-component dotted spinors with upper indices ” € F* and we have: Gi): Be (1.88) with : aa = A ¥() =b4e €C. (1.89) Hence ‘ Ope R= FM), and Per. Thus for dotted spinors we have to assign to }4 a row and to spinors with upper dotted indices columns. We then have: —B =z B wo b4= cist > and SS i lps By Wi =D. Furthermore, the correct description of the transition from F' to Fris given by complex conjugation and multiplication by the matrix!? 7° ()48 s FF So we set: (P)44(payt =. (1.90) The inverse map is found to be 0 wey (aa )* = da. (1.91) In agreement with Eq. (1.106a) (to be given later) we also have oA = Bo)", 19$ee Eq. (1.199) or Sec. 1.3.3 where the o-matrices are introduced1.3 SL(2,C), Dotted and Undotted Indices 43 and Baa vO) pa These relations show the connection between a and 9. ‘Thus the four complex numbers p,#" where A = 1,2; A = i,2, do in fact, define only four real independent numbers which we can take to be real ba, P4 where A = 1,2;4 =i,3. We now consider dual maps, From linear algebra we recall the following results. Let V and W be vector spaces over a field K, i.e. R or C. For every linear map @:V—+W we may construct the so-called dual map?° owt vt by the following prescription: Let ¢ W* so that # is a map b:W — K(=R,0). Then we define the dual map by $(v) = pod. This is to be understood in the following way. Applying the left hand side onto a vector v € V we have vev*® x, where ¢*(%) is by construction an element of V*, and this element maps v € V onto the real or complex numbers. The right hand side of this equation then gives: o:V—4W pew eV doy ew YEW. Hog eK. Hence the dual map ¢* is defined in such a way that the left and right hand sides of the equation give the same clement in the field K when applied to an element of the vector space V. Furthermore, a linear map ¢ then implies a linear dual map ¢*. It is important to note that if ¢ is a linear map from a vector space V to a vector space W then the dual map ¢* maps W* onto V*, i.e. dual ?°Here again we use the standard notation of linear algebra for dual spaces and dual maps which should not be confused with complex conjugation.44 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors maps have the ‘opposite’ direction. Describing ¢ and ¢* as matrices, we see that if ¢ corresponds to a matrix A, then the dual map ¢* is given by the transposed matrix A’. For a proof of this statement we refer to books on linear algebra (see for example [64]). With the help of the mathematical concept of a dual map we may explain the transformation properties of the various types of two-component Weyl spinors; within this formalism we find a natural explanation of Eqs. (1.60), (1.78), (1.62), and (1.81). We start by considering Eq. (1.60). Formally M:F—-F, ba > Ya = Ma? da, i.e. M is a map from the vector space F to the vector space F. According to the above formalism, the dual map is given by M-! : FY — F", M-!= M~'" as stated earlier (the dual map is represented by the transposed matrix). Here we choose the inverse of the dual map instead of the dual map itself as shown in Fig. 1.1. The figure shows that M~!" can be re-expressed in terms of a sequence of maps. We start with #4 € F*. Applying the matrix «4g we obtain the corresponding elements ya = eapy? of F. Then with p', = M,? pp we obtain the transformed element Y/, = My? yp = M,Pesoy® of F. Again applying the metric «4% we finally obtain the transformed element of F*. The result of this composition of maps has to be the same as applying M~!7 to the spinor y4 € F*. Hence we conclude (M717)4, = 48M, ecp. The map M : F —> F induces a corresponding map in the space F which can be constructed from M using the general prescription of complex conjugation, multiplication by the matrix 3°, and contracting with the metric ce! Starting with We =Ma?bn, wee F,1.3 SL(2,C), Dotted and Undotted Indices 45 Wy = MyFoe = MyPepoy F Figure 1.1: Directions of dual maps. we obtain the corresponding transformation in F"*: BA = 444)" = (0°)44(M a? bn)" = 6°)44(M4?) op = 44M) pe, using Eqs. (1.90) and (1.91). Contracting this equation with the metric ¢, 4, we obtain the transfor- mation in F, i.e. Bi- ead” = 48) 4(My?)(0) ped yP” = €jg(0") a (M4")" © poe Pen = €4p@)?4(My B)* (9) pee Ody, Defining ; : : ; - (M*) 4? = €4g@)*(Ma?)*(0) ge”,46 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors then : Ba = (MAD m and Mt: F — PF, in agreement with Eq. (1.62). The relation between the various spinors and their respective representation spaces is depicted in Fig. 1.2. wh = Balad PA = POO) a4 Figure 1.2: Weyl spinors and their respective representation spaces. Remark: For the matrix 1 2 Bye eae (m8) =(™ M5) © s112,0) the above definition of the matrix ((M*) ,”) implies:1.3 SL(2,€), Dotted and Undotted Indices 47 Note that the above unconventional definition of M* is due to our index convention; the connection between #4 and % plays a particular role. Again the transformation in F leads to a dual map in the dual space F'*. This corresponding dual map can be derived by contraction with the metric. Given io oS Ba = (M) bp and contracting both sides with e4¥, we obtain: ABB, = eee aM) 5 3p = Mar) 5° cone” Gp = a =: (unity 4 oe where a : D (MA = AC (Me ena, and MT; — F*. This is in agreement with Eq. (1.81). Again, for M € SL(2,C) with 1 2 Be ee (Ma?) = (ma ma) , the elements of the matrix M*~!" are given by: where In deriving the matrix M* we see that we cannot describe complex conjuga- tion as a linear map from F to F. This can, of course, also be shown explic- itly by considering a similarity transformation as in the previous equivalence proofs, and showing that an appropriate matrix does not exist.2! Therefore F and F define two inequivalent representation spaces for the special lin- ear group SL(2,C). On the other hand elements of the corresponding dual spaces F* and F* can be obtained with the metrics ¢4p and € 4g Tespectively. These are matrix representations of linear maps and therefore F and F* are equivalent representation spaces as well as F and F*. ?1 We shall see later that such a linear transformation exists in the four-dimensional Dirac formalism.48 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors 1.3.2 Calculations with Spinors We now consider a number of explicit calculations with spinors and derive several useful formulas. Proposition 1.13: Let } € F*, x € F, F being the representation space of SL(2,C). The quadratic form (x) = 1X (1.92a) is invariant under transformations of SL(2,C). Proof: We have to show that (Wx!) = (Wx), where w=M-'"h and y'=My. Consider (using Eqs. (1.82a) and (1.82b)) (Wx!) = Axa = (MO) Mao xe = 8 (M7) Mo xe =p bp°xc = 4? xB = (bx): Similarly Vax’* = bax’. Proposition 1.14: The quadratic form (OX) = Bax" (1.92b) is invariant under SL(2,€). Proof: We have (using Eqs. (1.82c) and (1.82d)) stay ay +) Bo s-1T)A oO OR) = Bax? = (Mi) goby (MOY 6x" = yt yB -it\A —é@_ +

H(2,C), cH» pla") = 2,0" _ [a =03 atic? (1.111) © \at ia? 2% +23 I1.3 SL(2,C), Dotted and Undotted Indices 57 where the four-vector is (2") 2°,x), (zz) = (2°,-x). Proposition 1,22: The map inverse to Eq. (1.111) pot: H(2,€) —> My is given by the following trace relation pt sHQ,€) — Ma, X esp \X)act= atk [xo]. (1.112) Proof: Using Eqs. (1.111) and (1.109a) we obtain: har [x04] = Lay [e,0"@) 7 it [o’a] 2, = 5 + Inlay = oh, Proposition 1.23: The determinant of p(x) € H(2,C) is given by: det X = det p(y) = a,c" = 2. (1.113) Proof: Using the definition of the map p, i.e. Eq. (1.111), we have: wag gly ‘S) al ig? 2423 det X = det p(z,) = det ( = (2°)? = (2)? = nwt” = 2°, The Hermiticity of the matrix X = p(z“) is seen immediately from the Hermiticity of the Pauli matrices X=xt=x*T, (1.114) We now consider the action of SL(2,C) on H(2, C), which is called the adjoint representation of the group SL(2,C). This representation is defined in the following way: ad: SL(2,C) —+ Aut(H(2,C)), M a: (1.115)58 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors with M' = adM(X) := MXMt, M,Mte SL(2,C), (1.116) where Aut(H(2,C)) is the automorphism group of H(2,C), which is isomor- phic to GL(H(2,€),C) [19]. We have to verify therefore that adM(X) € H(2,C), ie. we have to show that adM(X) is Hermitian. Using Eq. (1.116) we have [adM(X)]' = (MxMt)' = MxtMt = MXMt =adM(X), since X € H(2,C). From the adjoint representation (1.116) we can derive the index structure of the Pauli matrices, as indicated in Eq. (1.105). We had p(z") = X = 2,04, which transforms under the adjoint representation of SL(2,C) as: X'=MXMt, or ao" = Mo"a,M}. Using the index notation of the SL(2,C)-matrices M and Mt we conclude ao") 44 = Ma2(0") pav(M")? ;, (1.117) and we see that the Pauli matrices must carry a dotted and an undotted index. The matrices o# therefore map F into F, and similarly the matrices co” map F into F Next we calculate the determinant of a transformed Hermitian matrix. Using Eqs. (1.111) and (1.116) we have wia! = det(X') = det(MXM") = det(M) det (X) det(M') = det(X) = a,c", (1.118) since M € SL(2,C) the determinant of M equals 1. Thus the quadratic form 2," = det(X) is left invariant under transformations of the adjoint representation of SL(2,€). We now have the following construction. Starting from a Minkowski four- vector « € M4 we construct an Hermitian 2 x 2-matrix X € H(2,C). The determinant of this matrix is simply the Minkowski product of z, with itself.1.3 SL(2,C), Dotted and Undotted Indices 59 We then go to the adjoint representation of SL(2,C) on H(2,C€) and derive the result that the determinant is left invariant under SZ(2,C). If we apply the map p~! to the transformed Hermitian matrix, we obtain a four-vector z,, such that a,c!" = xy". We can demonstrate these steps schematically in the following way my 2 H(2,€) 24 H(2,€) 25 my, (1.119) that is: t4——> p(a,) = 2,0!» ad M——+ p(X") = «t, =X =MXmMt =X. This transformation from M4 to Mg is, of course, simply a Lorentz transforma- tion which transforms four-vectors into four-vectors leaving the Minkowski square, i.e. the line-element, invariant. Now from Eq. (1.1) we have al = Myx. On the other hand we can also write, using Eqs. (1.112), (1-116) and (1.111): att = Lay[x'o"] = br[Mxmtat] = p0y[Ma,o" Mio" = y[Mo' Mita", = 51 [o"Mo M2". (1.120) Comparing this expression with Eq. (1.1) above, we obtain: (1.121) Equation (1.121) is the explicit form of the group homomorphism SL(2,C) —+ Li. Proposition 1.24: The following properties hold: (i) Equation (1.121) is a group homomorphism, i.e. V Mi, M2 € SL(2,C) we have: AM, (M1)AY (M2) = AY (Mi M2), (1.122) and (Av*)M (4) = AFM"). (1.128)60 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors (ii) The matrix A, defined by Eq. (1.121), is an element of the restricted Lorentz group L., i.e. we have to show: det [A“,(M)] = +1 (1.124) and A°(M) >1. (1.125) Proof: We first prove Eq. (1.122). Using Eq. (1.121) we have a AY (M))AY (M2) = 7D [e"Mio,M{) Tr(a" MoopM}) = ED [Ma" mo.) 19" MzopM]] = 7 (Mtoe) 59" MopMiy; 1 F i o qleaton)”? (ov) ga )A4 [Moo pM3] 4 4 1 B ; 5 lion}? 56,4554 [Moo pM] 44 1 ar AA + a 5lMion Mi] [MropM3] 44 Tea ; = 5 [Miz*M.Meo,Ms)*; 1 - pi [Mia"M, Macy Mj) 1a eo = 5 B[a"( Me)o,(MiMe)*] = A¥(Mi M2), again using Eq. (1.121). A direct consequence of Eq. (1.122) is the following result: (A71)*)(M) = AH,(M~), (1.126) which we can prove as follows. Using Eqs. (1.122) and (1.109a): AM (MAY (Mo!) 2?) A (MM) = MAA sz0) fee (1.109a) 1 hayfateg] U2 3 ont, = 64, In order to prove Eq. (1.124) it suffices to recall that A*(Lsz2,c)) =1.3 SL(2,C), Dotted and Undotted Indices 61 and therefore det (A",(2s202,0))) = det (5",) = 1. Furthermore, det is a continuous function of A", so that det(A",(M)) must be +1 as M runs through SL(2,C). Finally, Eq. (1.125) may be shown as follows.?° Using Eq. (1.121) we have 1 1 A%(M) = 5 Tr[a°MooM!] = 5 Tr[MM"], since F = 0° = 09 = Nexo. Now let M € SL(2,€) and take U € U(2,€) such that UMUt =D, where D is diagonal. Then (Umut)! = Dt, or uMtut = Dt. Now 1 = det M = det(UMU) = det D = ayaa, where a1,a2 € C are the eigenvalues of M. Similarly 1 = det Mt = det(UM'U") = det Dt = aja}. Using the fact that a trace is invariant under similarity transformations we can write Tr[MMt) = Tr[UMM'Ut} = Tr[UMUtUMtU'] = Tr[DD'] = |a,|? + fag)’. ‘Taking into account that a) - az = af - a} = 1, we can write 4 1 T[MM1] = ar? + = ee la| las)? We now set = [af eR. 26We introduce a unitary matrix U which diagonalizes M.62 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Then: +1 j= St With this relation we can show that Tr MM? > 2. Suppose we have the opposite, i.e. Tr[Mut Tr[MM!] <2. Then : e+1 <2 c which implies (e-1)? <0. But this is impossible for any c € R,. Hence we conclude Tr[MM‘*] > 2, and hence AY)(M) = 51e(MMMt] > 1. Thus, A“,(M) as defined by Eq. (1.121) is, together with Eq. (1.124), a restricted Lorentz transformation. Our next step is to derive the map inverse to that of Eq. (1.121), i.e. given any Lorentz matrix A“, the question is: How can we construct the corresponding SLZ(2,C)-matrix M = M(A). We start by considering the action of the restricted Lorentz group on Minkowski space M4: cH= At’, chem, AEL. (1.127) However, we know that the action of SL(2,€) on Hermitian 2 x 2-matrices is given by the adjoint representation: X'=MXMt, MEéSL(2,C), X €H(2,C), or using relation (1.111): aio" =Mayo"Mt, ao, = Mx’o,Mt. Using Eq. (1.127), MY a’o, = Mx’o,Mt, so that Mo, = MoyM'. (1.128)1.3 SL(2,C), Dotted and Undotted Indices 63 Proposition 1.25: For every M € SL(2,C) o.Mo" = 2Tr[M] Lox2 (1.129) Proof: This can be shown by explicit verification. Muitiplying Eq. (1.128) by &” and summing over v, thereby using Eq. (1.129), we obtain for M, Mt € SL(2,C): Mo,0" = Mo, Mio" = M(2Tr[M]) 19,0, (1.130) and so M(A) = Moyo". 1 IMT Taking the determinant of Eq. (1.130) we obtain: det (A",040") = det{M(2TrM")] = det M - (2TrMt)? - det 12.2. Using det M = 1, since M € SL(2,C), we can express TrMt as a function of A, ive. f2 otret = +{det{A",0,0"] } ‘ The final expression for M as a function of A is therefore: (1.131) In components, Eq. (1.131) reads 1 : M4? (A) = +-————p M (oy) an(0")”?. {det [A",0,0°]} Thus the Lorentz indices y,v are effectively exchanged via 0,7 for spinor indices A, B. Equations (1.121) and (1.131) show the connection between the two groups SL(2,C) and Li. An important point is that both SL(2,C)-matrices M and —M lead to the same Lorentz transformation as can be seen from formula (1.121). In Eq. (1.131) this fact corresponds to the + signs. Thus the correspondence A <> +M defines a two-valued representation of the64 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors restriced Lorentz group, which is called the spinor representation. This leads us to the identification Li, & SL(2,€)/Z2, (1.132) and so SL(2,C) is the universal covering group of L', which means the SL(2,€) is simply connected [104]. 1.3.4 The Fierz-Reordering Formula The Fierz reordering formula exchanges the order of anticommuting spinors. Proposition 1.26: The following relation holds for spinors 7%, ¢, x: i (GP)X 4 = (GO? R) (Hou) 4» (1.133) Proof: Using Eqs. (1.92a), (1.84), (1.86) we find: (OW)xX4 =o? bBX, = Peace jpx? = -ecng eye gpx” = cay Pei px° = 6405 MP ecneg py e?X”. Using Eq. (1.110) and finally Eq. (1.106a) we arrive at: 1 : (GW)X4 = 300) Aly Cecre nny oP? il * BH i = yl) ie" Op) py ecnen nV OX” i : = 50") salu) eH” Bd" pwroPX? 1 i = 59 (on aoX Yo") a4 1 ~ = 4 (404%) (0%) For further Fierz reordering formulas see Eqs. (1-101), (6.90), (6.91), and (6.92).1.3 SL(2,C), Dotted and Undotted Indices 65 1.3.5 Further Calculations with Spinors We begin by deriving some more formulas which are very useful in calcula- tions with spinors. Proposition 1.27: Let ¢ and x be spinors and o,a” given by Eqs. (1.105) and (1.106a), then the following identiy holds: (¢0%X) = —(xo"#) (1.134) Proof: We start with the expression on the left: ) (1928) 4 89) (AB bey M 2 Me Me : (¢0"X) ia (0%) ax? ta on(o") spe Xe C3 1,106: | fa = —bpeP (alt). pe®oXe | ad ~op(-a") PX 2) x(a") P by = -(Za"6). Proposition 1.28: Under Hermitian conjugation the binary spinor expressions ¢o"X and (@¢) transform according to: (¢o"%)! = (xo"), (1.135a) (69)' =6¢. (1.135b) Proof: The first expression, Eq. (1.135a), is $L(2,C) covariant. There- fore, with the definition of the {-operation (Got)! = (xtoMt gt) = xIA(0Ht gt? = x4(oM) 4a = (x08). Here we use: VAr= yt and gi8 = 8". It may be observed that the Grassmann property, Eqs. (1.94) and (1.95), has not been used because (¢x)! is defined as Xt ¢t. Similarly, Eq. (1.135b) can be demonstrated as follows: (66) °2 (gat =o1,'? = 35. From Eq. (1.135a) we see that yo“ is Hermitian.66 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Proposition 1.29: As can be expected, the expression (¢o/X) transforms as a four- vector under the Lorentz group, i.e. (¢'o"X’) = A",(M) ($o"X)- (1.136) Proof: We have: (Worx) C2) — xara") = -x4(o")4? oy ~(M*) Px p(o")4? Mp Coc = -X_(Mtt)? (4? My de = ~Xa" (MN? 5(0")4P My Ec bv —F¥HP Corea MI)? ("PM yb E males i(o")" Mp“(ov)ea] (Ra(@")?? 40) a -har[Mi4 Mor] (xa") = A¥,(M)($o"X), using Eqs. (1.121) and (1.134). Proposition 1.30: The following relations can be shown to hold: ge = (oP aa? (608) = ba (08), (1.137) (80¥6)(00%8) = jn” (80) (00). Proof: Using Eqs. (1.92a) and (1.92b) we have (808) (608) O4(o!) 498 (oop = -04(04) 30°" (0 o30” a = -(0") 45(0"\op0400?0” a Flo aalo"ope*eP? (00) (80) Co i ) gn(@")°4 (06) (88) oo ite 2).1.3 SL(2,C), Dotted and Undotted Indices 67 Proposition 1.31: The generators of the group SL(2,C) in the spinor representa- tions (1/2,0) and (0,1/2) are given by (see Eq. (1.131)) ol = Floto - ova"), (1.138a) at = =(GHa" — aa"), (1.138b) (om) t = oH, (1.138c) We will need these two-dimensional spinor representations of the Lorentz generators later for the construction of the complex two-dimensional grading of the Poincaré algebra. In components, Eqs. (1.138a) and (1.138b) read: ft tie A a (ot) 4? = Z[(o") 4a(0")*? = (0) 440")? ], (ay, = L@AA(0") 4p — YAO!) a5) Proof of Eq. (1.138a): For an infinitesimal Lorentz transformation (w;; infinitesimal) we have: Moya” = Kyoto” 1 wor Woz wos _—[-eo -1 wi wis bn 7 ote wp. —wi2 —1 wg —w3 wig ang —1/ 4 — 2(wo3 + iwy2) —2(wo1 + twas) +2i(wo2 + iws1) —2(wo1 + twas) —2i(wo2 + iwsi) 4 + 2(wo3 + twa) Hence: det (A“,0,0") = 16 + Ow’), and, omitting terms of order w?: He N= [det(a",0,0")| = The relationship between an element A’, of the Lie group SO(1,3;R) and an element M“” of the Lie algebra so(1,3;R) is AP, (w) = [exr(-juum”)]"68 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors For infinitesimal parameters w,,, the element M(A) € SL(2,C) is M(A) = FM lwo" = 55 (tor ~ fee M! oP = ven = Seat, where 1 aM t= py (MM) 0°97. Now 2 = Mgr” = (pr + Wpr)2", and since every homogeneous linear transformation group possesses the triv- ial self-representation in the four-space spanned by the z,, we have: ww (MM) = Wor. This equation implies for the M’s: (ery? = i (nt a”, — nPPdH,). Of course, from Eq. (1.11) we know that the generators of the Lorentz group have this form. Hence a se a a ey at = 4. (tes, — #8, opa? = ¥ (ator — oot). In a similar way, a4” is obtained by considering M* as a function of the Lorentz transformation A. The matrices 0” and @” obey the commutation relations of the Lorentz algebra so(1,3;R), i.e. Eq. (1.14). Proposition 1.32: The generators of rotations Ji = 1,2,3, and Lorentz boosts K;,j =1,2,3, defined in Eqs. (1.19) and (1.20), i.e. - i= yee, Ki = Mo can be written: = 7 j a F = SA yo = Sota, (1.139a) (1.139)1.3 SL(2,C), Dotted and Undotted Indices 69 Proof: pee ‘i oF eel & Fe = Seigno® = Sein (o93* — oFat) WY) Hein. (—oFoh + ota!) i i 7 1 1 iE gcse ila, = zenilon = 3h = 9% = — ein [07,04] where we used 3 DY egee™ = 26. jk=1 Relation (1.139b) is shown as follows: K;=0%= 5 (oa -o'@°) where we used again Eq. (1.108). Proposition 1.33: The o” are selfdual, i.e. ol = Jet? ope, (1.140a) whereas the 4” are anti-selfdual, i.e. ay = ~ peo pn: (1.140b) Proof: In order to prove the first of these relations consider (i) p=0,v = i;4 = 1,2,3: t 1.108) @ oP! = £ (oa! — ata?) OY) 4 4 On the other hand with i, j,k = 1,2,3: 1 1 lt ipa: a ijk Ee Oijk pee Hae Opa = FE Ojk = we Tos, OKO. 2i po = gg Moje = 30g (Ose — OH i) = go (290 +0403) = — per [o5, 04) 04; t 1 vise, vl = — = tsk; = — Xk Ce te jKIe 0, we obtain: hy OuUw = 0, (1.167a) which can be written, using Eqs. (1.159) and (1.153), (ae ("2 a0) ($3) ae and therefore the 4 x 4 matrix equation (1.166b) decouples into two 2 x 2 matrix equations: i(o"d,) > — mo = 0, (1.167b) i(G*0,) 4 — mp = 0, (1.167c) i.e. for m — 0 we have: a ARG, oy 6A, al g)Be =(8)48 (4 2) pe + 458, ~ (073) 46 (0) 4g - (0) ge “tape, (@) ac -(@0?)4, +64, Using a” as matrices and (0?)? = 1x2, we obtain: Seco 0 vat} 0 ah 4 gd] = Bae Cc ¢, We now investigate the individual y-matrices. p=0: (Re) aa = Yab(Y) se Yea" 1 64 (07)an\ (58° 0 2 (tei i. ( 9 »,) ( BP + (0%) sal)? (0) an + (*)av -3( —(g2)4P AD ~(8) 46 (92), | 5A, ) “(aie M4) -Co 2) again using Eq. (1.108), i.e. ? if 8% © )as\( 9 (o')ee\( 5c? we) “3 \_(g2)4B -A, (eye 0 ~(#)¢ 7 = -o? and (07)? = Lox2. (0?) 49(#1)8? ee (ae 7 (we? —(a')AD +. (6)4B(a1) peg ag (0?) 49 )°°(0 op ~ ("an ) Oop + PMO) ps98 CHAPTER 1 Lorentz and Poincaré Group, SL(2,€) and Spinors 1 ( (07a) 4? — (070?) 4? (0?a10") 4p - ae) <2 =(@)4? + ole’ g)Ad -(@0)4 + (#0), mle (o!0? — o?a!) 4? tet ate} (-# a (0!0? — 070) 2i(o5) 4 -(ealo? +0) ap (o 8 oe 2i(o)A 5 vin D 1 a) oe (on a) a 0 ) oe ean) Ale 2 re aA 2i(0°)4, 0 io), 1( &? Jas 0 (*)be fe? (ew 9 —(#2)48 64, (eae 0 — (REP 5 io (0 2) AB : (wybD — (e8 Vago? . (© (#2)42 + (22/48 (0?) palo?) (07) 43)? ep ae | =)? op + (49 (07) pp, 1 0 -(0? = 0°30?) 45 a9! 2 — 5292G2)AD 0 ) Again using Eq. (1.108), we obtain 2x2 -0? - 0°a?o? = 0°(0?)? = 0°, since (02)? = 1x2. Therefore: 1 0 —2(07) 0-0? (dai (aero 0") = (et) pak (Re) aa = Yao 7b) y-Yea™ oe bas (0) a8 0 (0%)pe a7) —(62)48 -A, (wy Be 0 ee1.4 Dirac and Majorana Spinors 99 u i (0764) 4? ~ (0332). (00) 49 - (0) ay ) By (BAP 4 (G%o3q2)4D ~(380?)4, + (#o8)4, aa (mene oat ) 2 0 -((07,04)*5, _ 1 f-2i(o') 4? 0 2 0 -2i(0')4 _ Ce 0 ) = va 0 alder Now the connection between the Weyl representation and the Dirac rep- resentation is given by Eq. (1.173a), Tp =X7'TyX. The Dirac representation and the Majorana representation are related via Eq. (1.180), Ty =YPpY"l. Thus Tu =YX'TwXxY", or, using Y = Y~!, we obtain the similarity transformation that connects the Wey] representation and the Majorana representation: Ty = (XY)7'Pw(XY). (1.182) With L (-42x2 0° ) —1\T -1yt oo =(x-')7 =(x7y, 3 (7? 2) ey = 0) de ioyo ae + -1 Y=— =Ytey ee i) a we obtain: tae Gap 53° (one XarYoe=3\ _coyas _ 5A gyic _gb 2 \-(a) -54, } \-@) -58, 1 (548559 ~ 0) 4n@)P° -447(0") pe ~ (0°) 458% 2 \ -(99)4B 5,6 + 64, (5?)P° -(59)4B(G?) gg + 04,55", ey sitar +(07)4° (0) 46 - (ae ~ 2\ _(@yde 4 eae ~(0?)A, +54, :100 CHAPTER1 Lorentz and Poincaré Group, SL(2,C) and Spinors which is usually written in matrix form as: 1 (-1+0? -o?-1 xy=3 (47% 751): (1.183) The inverse, (XY)~!, is a eT lft o?\ /-1 -o® loy-ty-be (XY) YUCX z ( 2 ) ( 0 4) +077 —o° a Fo + :) > (1.184) which is usually written -1_1/-1+4+067? -1-¢? t => = ) (xY) z (he 1 0? 41 (xy), (1.185) or (xy) =y-'x-t = ytxt = (xy), However, for a clear index structure it seems to be advantageous to use the explicit form Eq. (1.184). Of course, Eq. (1.185) is the same matrix as Eq. (1.184) since as matrices: F = loxe, Ce Oa and so on. Next we check that he = (XY) My (XY) 0 0 o Ww=\x o}° We perform this calculation in the compact matrix form of Eq. (1.185): 7 1 (-1+0? -1-07\ (0 1 2p? = 1 (XY WY) = 3 (Ch -1 4) ( 6) (2 -o? 1-0? ) _1((-1 = 07)(o? - 1) + (-14 9?)(-1 - 07) 74 (1 - 0?)(o? - 1) + (0? +1)? —(1 + 0?)(-0? — 1) + (-1+0°)(1 - 0?) (1 - 0?)(-0? - 1) + (-0? - 1)(1 ~ 0”) _1(0 40%) _ (0 o\_ 19 fade 0 eo 0) In the last step we made use of Eq. (1.178). The other 7-matrices can be checked in a similar fashion. with1.4 Dirac and Majorana Spinors 101 1.4.4 Charge Conjugation, Dirac and Weyl Representations The charge conjugation matrix appears in Dirac theory*? in the following way. The Dirac theory implies the existence of electrons and positrons, par- ticles with the same mass but opposite charges, which obey the same equa- tion. The Dirac equation must therefore admit a symmetry corresponding to the interchange of particles and antiparticles. We thus seek a transformation vo YP, which reverses the sign of the charge, so that the Dirac spinor Y obeys the Dirac equation (:@-eA-m)¥=0, A= Ayr", (1.186) in the presence of the electromagnetic vector potential A,, whereas the charge conjugated spinor Y° obeys*> (i P+eA—m)¥°=0. (1.187) ‘The Dirac equation coupled minimally to the electromagnetic field is [7 (id, — eA,) — m]¥ = 0. Taking the complex conjugate we obtain: [y"* (id, — €A,) — m]¥* = 0. ‘Transposition yields WL (i By ~eAn) ~ mi] = 0, Wty? Pitty? (i By -eAy) Wy" (-i dn ~eAy) — m] = 0, where is the Dirac adjoint, and Potty? = yf. ‘Taking again the transpose we obtain =T [-7"T (i8, + eA,) — mW = 0. See e.g. the discussions in A. Pais [86], p. 381, and J.D. Bjorken and S.D. Drell [17]. *8For further discussions of these aspects see M.D. Scadron [103], Chap. 3.B.102. CHAPTER 1 Lorentz and Poincaré Group, SL({2,C) and Spinors Multiplying this equation by a 4 x 4-matrix C from the left and inserting C-'C in front of U", we get: C[-1"" (id) + e4,) — mlO“'0" =0, [-Cy""O"! (18, + eA,) — m|CW" =0. This equation can be identified with Eq. (1.187) provided we set: w= C0'," (1.188) except for a phase factor. In addition we have to demand that in any repre- sentation of the y-matrices CyTOnh = 7". (1.189) The matrix C is called the charge conjugation matriz. It suffices to construct the charge conjugation matrix in some particular representation; the unitary transformation which transforms to another representation then gives the matrix C’ in this new representation. We consider several representations. The Charge Conjugation Matrix in the Dirac Representation In the Dirac representation (1.56) the charge conjugation matrix C may be taken as j 0 oe Cp = inn =i (@ a ) . (1.190) Proposition 1.46: The charge conjugation matrix Cp defined in Eq. (1.190) pos- sesses the following properties: Cp = -C5! = -ch, = -CB. (1.191) Proof: We first verify Cp = —C5. Consider ara eRe —(Co)ab(C) be e a (caryee “r) tl I eee: aL i o1.4 Dirac and Majorana Spinors 103 (1.108) (a 0 ) OG. This shows that Cp = —C>}. Next we evaluate the Hermitian conjugate as follows: aay Nt gt (Chu =|3( rae ()aa\| =i . Mare aa (=) 9 (07) 0 0 -(0%) a5 “ (cosie 7) se using the Hermiticity of the Pauli matrices and ? = —o? from Eq, (1.108). Finally, the antisymmetry of the charge conjugation matrix Cp is shown as follows: 0 -(%)4,\'_.f 0 27) 45 (Co) ay =# (corpse oe) rt (cerryae : ee) 0 seat a 0 2 = (geryis ")*) = (Lierae ° 0) 0 fie at’ at (cosy | = ~(Co) ay Here we made use of Eq. (1.108) and the antisymmetry of the 0? matrix. Proposition 1.47: The charge conjugation matrix Cp, defined in Eq. (1.190), sat- isfies relation (1.189), i.e. Cop! Cp! = Yh. Proof: First we attach suffixes D to Eq. (1.189) into the following form. Corb! Op! =~}. Using Eq. (1.191) this can be written as . -CpYp' Cb = ~Y-104. CHAPTER 1 Lorentz and Poincaré Group, SL({2,C) and Spinors Transposition gives . ChyYBCD = Yb» and using the properties (1.191) again yields CpypCo = yee Consider the case ps = 0, i.e. (Cp)a(Yb)be(Co)ea = Next we work out the case y= 1: 1 : Cle1.4 Dirac and Majorana Spinors 105 Using Eq. (1.190) we have: 0 -(0%),; 0 ives (Co)an(rb)te(Cr)et = — (capa “ae (apse 7 je) 0 (92). : (een 7) : -( oe aslo" (2)49(0!)go(o")?? 0 a -( 0 : ne) ae (@ala?)AP 0 = 0 —(0?010?) 45 (o?at92)AP 0 a 7 =- (Gave (o .*”) (cate? = -o') ease _tityia & 40) (0 = ot) = (Dea since 0 qe 0 ity. Oa (piso ao = ((mao i 0”) 0 ("ap ee (#T)Ad 0 : Hence we have demonstrated that CorbCr = (1b) Next we evaluate the case ps = 2: 0 2: i (bas= (gayi 9M) Using Eq. (1.190): 0 ~(0?) ag 0 (0)b6 (Cp)av(Yb)te(Cd)ea = - (geo 0 (ease 0 ) tee“)106 CHAPTER1 Lorentz and Poincaré Group, $L(2,C) and Spinors {0 (2) 48° cp (3?)49 (0?) ge(@V°P 0 . 0 (078? o) 45 (F02a2)AD 0 = (gran a : Now the transpose of 77, is: 0 2) .\T 0 elie GoD aa we = ((aro a 0”) Then: oe 0 ee (Co)aalrb)te(Co)ea =~ oC. ae -e ae so (gs40 ] Now 0 ayn 0 (2) aa OB Yaa = (asy4o a 0) = (ary : ) (aie 4) (atin1.4 Dirac and Majorana Spinors 107 using Eq. (1.108) and 0% = 037. Hence . CoybCp = (yb) | as had to be shown. The Charge Conjugation Matrix in the Weyl Representation We transform the charge conjugation matrix Eq. (1.190) from the Dirac rep- resentation to the Weyl representation by using Eqs. (1.173a) and (1.173b); thus Cw = XCpX7! =iXyh7hXx7! ary = a i) CC) -io’ = iXyp X71 X49) x7! We prove Eq. (1.192) in the submatrix formulation, avoiding cumbersome indices. Thus Cw=i(-t °°) (9 -2\ (-1 -0° WE 2\-09 -1)\o? 0 J\O -1 i (Goze - 0°? -o? — aie) © 2100? +e? ao? + Fo" et (2on 08) (doe 0 ~2\0 -207) ~\ 0 -io?}* Remark: The correct form of the charge conjugation matrix in the Weyl representation, possessing the correct index structure, is 250 to*e 0 Cw = ( ; ia’) (1.193) with: 7 (Cw) i(0?a®) , 0 W) ab = fe . 0 i(a*o°)*, Of course, the matrix (1.193) is the same as Eq. (1.192), since 3° and o° are unit matrices, which are usually ignored in the literature. Proposition 1.48: The charge conjugation matrix Cy in the Weyl representation also satisfies Cw = -Cy} = -Ci) = -ch,, (1.194)108 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors and Cwrty Cy = —y. (1.195) Proof: Equations (1.194) and (1.195) can be checked directly, using Eq. (1.192) and Eq. (1.159), or alternatively with the help of Eqs. (1.173a) and (1.173b). Thus using first Eq. (1.192) and then Eq. (1.191) we have: Cw = XCpX~! = ~XC05!X7! = —(XOpX"') | = CF}. Also Cw OS? xcpx7) 2? xh xe! = (x7 ep x") * (1.182) ~(X@pX7!)* (2.192) -cy. The last property in Eq. (1.194) is shown as follows: Cw = XCpX = -Xchx" = -(x~Hopxtyt = —(XCpx7')' = -chy. Finally we verify Bq. (1.195). We know that in the Dirac representation we have: Coy,C5! = -78"- Hence: CorbCp' = —Ibs = X(CovgCp)X-1 = -X (76) X73, = XCpX7XyhX-'XCGIX7) = -X (7p) XT, 3 XCpX1 XY XA(XCpX' | = — (XoXo) and so ee T Cw Cay = -Hy - Further properties of the C-matrix Proposition 1.49: Independently of any representation, the charge conjugation ma- trix satisfies the following relations:14 Dirac and Majorana Spinors 109 CYC! = (°)", (1.196) O(ryH}O7} = (Py#)" (1.197) Proof: We first show Eq. (1.196). Thus, since Eq. (1.162a) is independent of the particular representation we have: oC =ic Py pot =i0yC"Cy'o" cro“ oxo" = if yt pst ( with (1.184)) = i(Py2y'y)™ = i(y yy)" 7 =(%)', using the basic Clifford algebra relation (1.160), which is valid in any repre- sentation. Equation (1.197) can be shown in a similar way: CoP yH#O7! = CCE HOM = = (9) (4) = -(9"')" = (oT with Eq. (1.164). 1.4.5 Majorana Spinors We consider a Dirac four-spinor in the Wey] representation. According to Eq. (1-156) we have: _(¢ ba . W= (§).- a ) eFoF"*. (1.198) For the index calculus to make sense, rows and columns must have the index structure of (i), 3), amt w),= (52) respectively. Thus (es us) és) _ My?d0+ Mygb” MAB MA, v M4865 4 MAD”110 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors and M, Map ~ ; eae ace (#4, wie ms) = (94mg? +B? 94M, 8 + Bam). B. We now observe that the relationship between left-handed and right-handed Weyl spinors may be written = PHO )PA, 84 = 80). (1.199) ‘The consistency of these relations with Eq. (1.63), i.e. wad, (1.200) follows from the fact that the 0° and @° matrices are 2 x 2 unit matrices. It may be noted that 44 is a two-component column vector, whereas WA isa two-component row; similarly P' is a column and a is a row. The consistency of Eqs. (1.199) and (1.200) can also be seen by calculat- ing the Dirac conjugate of W, i.e. Using Eq. (1.200) we have: Tw == (4s BD (so, B°) = Gi =, 34): Note that since Eq. (1.200) does not preserve a consistent index structure the matrix representation of 7° must be used. On the other hand, using Eq. (1.199) we have: a 0 0) Ww = Vip = (64%, Ba) ae te 38) 1.199) - = (Gi@) AB, g4*(0°) 4 A) OL (YF, by (1.201a) in agreement with the previous result. We note that Ty = (5) : (1.201b) The charge conjugate of a Dirac spinor in the Weyl representation is defined as in Eq. (1.188) with the charge conjugation matrix Eq. (1. 192), te. Sy = CwTw.1.4 Dirac and Majorana Spinors M1 Thus using Eq. (1.193) and Eqs. (1.201a), (1.201b), we have: (Yi), = (Cw) aT), ' (eye ay (3) (7 mt 3) Now (irony? = ( and (i@o°) ig = ( Hence (io?a°) 4? = erc(io?a®)? = eace®? = 54°, and (ia?o°)* pa 4° (igo) =e ea = Hence: (Viv), = (5) (1.202) Thus, charge conjugation flips ¢ and y. A Majorana spinor is a four-component Dirac spinor which satisfies (here in the Weyl representation) Uy = Vy. (1.203) Using Eq. (1.202) and the explicit form of a Dirac spinor in the Weyl repre- sentation, i.e. Eq. (1.156), we obtain eee) ee o e ‘Thus for a Majorana spinor in the Weyl representation we can write: wi = (5) . (1.204) Thus a Majorana spinor has only two independent complex components and is therefore equivalent to a two-component Weyl spinor or a real Dirac spinor.112. CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors 1.4.6 Calculations with Dirac Spinors It is useful to know the connection between the four-component Dirac for- malism and the two-component Weyl formalism. The use of the Weyl repre- sentation of the four-component Dirac formalism has certain advantages. We use the following notation for Dirac spinors: ba w=("), 1.205; (%) 2 where the subscripts + distinguish between Weyl spinors which are elements of representation spaces F (+) and F* (—) respectively. Then according to Eq. (1.201a), the Dirac conjugate of Eq. (1.205) is: =(¥4, D4). (1.206) Proposition 1.50: The following relations hold for Dirac spinors Y and y: (Ux), = (W-x+) 2 + 4X-)os (1.207a) (Bx), = —(H-x+)2 + PX Jo» (1.207b) (Y'x), = ©. met (v-0"X_) 5, (1.207e) (Fy? x) = (d-0" XJ — G45" x4) 95 (1.2074) (Woy), = (v-o8' “a+ (B74 R Jos (1.207e) where : of =o [yt]. (1.207f) Proof: (a) On using Eqs. (1.92a), (1.92b), and (1.93), we obtain for Eq. (1.207a): (Fa), = (4, Baa) (54) = vera + Bat = (b-x+)o + (OX) o- (b) With Eqs. (1.92a), (1.92b), and (1.93), we can prove Eq. (1.207b): Tye T.45 A moa 0 (xe (GPx) 4 = Vortoxe = (44, Ya) je 4 a B e = (04, Ba) (2G) —Whxea + Bad = —(Y-x4)2 + (BX )o»14 Dirac and Majorana Spinors 113 (c) Proof of Eq, (1.207¢): a 0 "ap (Gtx), = Vorkxe =” (WA B a) Ce 7 7) (S ) (nyA 0 = (47a) (Cag ) PX B = V0!) sax! + D4 Pxee = (bo F_)y + G.T"x4)o» again using Eqs. (1.92a), (1.92b), and (1.93). (d) Proof of Eq. (1.207d): (Go' x) 4 = Va Vea teXe : 0 (o#)4an\ (~o2° 9) (xse = (v4, ¥,4) (ee u) ( 0 #,) (<2) “a 0 a a = (#2, 44) (cone 7 aK 3) . *) axe = (4, B,) (“ oes ) = V8(0") 5X2 - Bia) xee = (pot X_)y = (GF x4) a» again using Eqs. (1.92a), (1.92b), and (1.93). (e) Proof of Eq. (1.207e): Before we demonstrate Eq. (1.207e), we discuss Eq, (1.2078), ie. ("y= G(71),,= *(otent, ~ hers) (gic OMe (amyae “) (ere ")114. CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors i oe y's) using Eqs. (1.138a) and (1.138b). With this expression one can directly verify Eq. (1.207e): aa o! 0 Won), (8, (MH cana,) (A) = JaPX+B = (v4, Bea) (oe wg) = VA (oh) aP xia + B44 gee = (H-08"x4)9 + (G472"X-)o» using Eqs. (1.92a), (1.92b), and (1.93). 1.4.7 Calculations with Majorana Spinors As explained in Sec. 1.4.5, a Majorana spinor has the characteristic property that we can replace the dotted Weyl spinor by the complex conjugate of the undotted one. In the terminology of the previous section this means that we can replace y4 by v4 and @° by Bs. Thus we can drop the suffixes + in Eq. (1.205) and write Un= (3!) i (1.208) In this case, Eqs. (1.207a) to (1.207e) read: (Gurxar) 4 = (bx)2 + (PX)o> (1.209a) (GarPx), = —(4x)2 + (PRos (1.209b) (Gur'xur), = GF"x+) 2 + (YOR) »» (1.209c) (Emr'Pxar), = (¥o"®X)_ — (Fe"x)y, (1.2094) (Guo’xm) 4 = (Yos"x)y + (072). (1.209e) Proposition 1.51: The following relation holds for Majorana spinors: (Gaerty ar) (Em PWa) = 0! (Cua) (1.210)1.4 Dirac and Majorana Spinors 115 Proof: Using Eq. (1.2094) we have: (Eu) (Eure va) (1.2094) [- Ge"), + (vor) | [- Go"), 7 (¥o"),| 29 [(otD), + (o").] [(Uo"D), + (Wo"),| = 4(Yo"B),(¥o"'B), OE” 20h” (W¥) 9B D>. For the present case Eq. (1.209a) is (Waa), = (¥%)o+ (b4),. Taking the square of this expression we obtain: (Gar ¥an),)? = (WY). + GV)” = (WH) 2+ (WO) 2 + %).(0¥) 2 + (FV) .(0%)y = 204) (FF), since as will be shown below: (Ve) (0%). = (UD) o(FH)y = 9, and (W¥)20¥)2 = PVo(WY)o- Hence it follows that (Gu Var) Gary’ Var) = 0” (Ea Yan) §- We now show explicitly that (y)3 = 0: (¥)2 = (H¥), (2), O29 (yAga) (vn) 039) Ao cpac doen = ACP bobadodn = Ppp bat + 72 yadda te dodo + edi dedi de = vovrdodr — dod dive — dridedati + didedide-116 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Taking into account the fact that #4 are Grassmann numbers, we obtain (cf. discussion following Eq. (1.96)) (wo); = 41d vod = 0, since for Grassmann numbers 4) {vv} =0 implies biti + bith = 0, ie th = A similar calculation can be carried out for (0%);. We now verify that (B¥)_(¥d), = (YH) o(¥9).- To this end consider PV a(00), = Fil Ws = Ta Tn = VDT ba = Vdd O" = (¥0), (OV We prove one more formula. Proposition 1.52: The following relation holds for Majorana spinors: ¥ 2g 2 (Gu Uo), = —- Gua) (1.211) Proof: Using Eq. (1.209b) we have: (Futa)3 = {-)a# +@B} = (WW)} - 204) (B9)_ + FD = -2(U0), (09), = -{(6v)3 + 2(0v) (8). + G3} -(Gu¥m)y, where in the last step we used Eq. (1.209a).Chapter 2 No-Go Theorems and Graded Lie Algebras 2.1 The Theorems of Coleman-Mandula and Haag, Lopuszanski, Sohnius We now discuss the two theorems already referred to in the Introduction.! 2.1.1 The Theorem of Coleman—Mandula ‘The following theorem? was established by S. Coleman and J. Mandula [23]: Let G be a connected symmetry group of the S-matrix, ie. a group whose generators commute with the S-matrix, and make the following five assumptions: (i) Lorentz invariance: G contains a subgroup which is locally isomorphic to the Poincaré group. (ii) Particle finiteness: All particle types correspond to positive- energy representations of the Poincaré group. For any finite mass M, there is only a finite number of particles with mass less than M. 14.0. Barut and R. Raczka [8], p. 43, also give considerations concerning the unifica- tion of the Poincaré algebra with internal symmetry algebras. Stronger group theoretical lemmas are given on p. 629, and it is pointed out (p. 630) that the infinite-parameter Lie algebra associated with noncompact dynamical groups (which lead to infinite particle multiplets (cf. pp. 411, 609)) does not contradict these theorems. See also the discussion in S. Weinberg [118] : 7118 CHAPTER 2 No-Go Theorems and Graded Lie Algebras (iii) Weak elastic analyticity: Elastic scattering amplitudes are analytic functions of centre-of-mass energy squared s and invariant momentum transfer squared ¢ in some neighbour- hood of the physical region, except at normal thresholds. (iv) Occurrence of scattering: Let |p) and |p’) be any two one- particle momentum eigenstates, and let |p,p') be the two- particle state constructed from these. Then T\p,p') # 0, where 7 is the T-matrix defined by $= 1-6(2n)*6'(py - pT except, perhaps, for certain isolated values of $. In simpler terms this assumption means: Two plane waves scatter at almost any energy. (v, Technical assumption: The generators of G, considered as integral operators in momentum space, have distributions for their kernels. Then the group G is locally isomorphic to the direct product of a compact symmetry group and the Poincaré group. We recall briefly some basic results of scattering theory. The Hilbert space H is the direct sum of an infinite number of subspaces, #.e. H=MM OHM. Here H denotes the n-particle subspace. It is a subspace of the direct product (symmetric or antisymmetric in accordance with the generalized exclusion principle) of n Hilbert spaces, each being isomorphic to HW, The S-matrix is a unitary operator on H. A unitary operator U on # is said to be a symmetry transformation of the S-matrix if (i) U transforms one-particle states into one-particle states, (ii) U acts on many-particle states as if they were tensor products of one- particle states, (iii) U commutes with S.