Representation Theory of Classical Compact Lie Groups Dal S. Yu Senior Thesis Presented to the College of Sciences The University of Texas at San Antonio May 8, 2011 i Dedicated to my family, and to Kim Hyun Hwa ii Acknowledgements First and foremost, I would like to thank Dr. Eduardo Duen˜ez for his mentorship, inspi- ration, and unparalleled enthusiasm. Because of him I have a much deeper appreciation for mathematics. I wish to thank my thesis readers Dr. Manuel Berrioza´bal and Dr. Dmitry Gokhman for making this thesis possible. Special thanks to Dr. Berrioza´bal for his infinite wisdom and encouragement throughout my undergraduate years. I also want to give thanks to both sets of my parents for their support, especially to my stepfather, Keith. Contents 1 Introduction 1 2 Bilinear Forms 4 2.1 Bilinear and Sesquilinear Forms . . . . . . . . . . . . . . . . . . . . . . . . . 4 3 Classical Lie Groups 8 3.1 The Orthogonal Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 3.2 The Unitary Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.3 The Symplectic Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.4 Lie Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.5 Lie Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 4 Elementary Representation Theory 24 4.1 The Normalized Haar Integral on a Compact Group . . . . . . . . . . . . . . 24 4.2 Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 4.3 Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 5 Representations of Tori 36 iii CONTENTS iv 5.1 Representations of Tori . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 6 Maximal Tori 40 6.1 Maximal Tori . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 6.2 The Weyl Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 7 Roots and Weights 47 7.1 The Stiefel Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 7.2 The Affine Weyl Group and the Fundamental Group . . . . . . . . . . . . . 50 7.3 Dual Space of the Lie algebra of a Maximal Torus . . . . . . . . . . . . . . . 52 8 Representation Theory of the Classical Lie Groups 54 8.1 Representation Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Chapter 1 Introduction In this thesis, we will study the representation theory of compact Lie groups, emphasizing the case of the classical compact groups (namely the groups of orthogonal, unitary and sym- plecticmatricesofagivensize). Theseact(are“represented”)asgroupsoflinearsymmetries endowing vector spaces with extra structure. Such vector spaces are called representations of the respective group and can be decomposed into subspaces (irreducible representations). Oneofourmaingoalsistoexplaintheabstractclassificationoftheirreduciblerepresentation of classical compact Lie groups. Although we will not have the opportunity to explore applications of this theory, we must mention that Lie groups and their representation theory can be found in many places in mathematics and physics. For example, the decomposition of the natural representation of the special orthogonal group SO(2) on the space of (say, square-integrable) functions on the circle (or, what amounts to the same, periodic functions on the line), corresponds to the Fourier series decomposition of such functions, while the irreducible representations of SO(3) 1 CHAPTER 1. INTRODUCTION 2 on the space of functions on R3 correspond to the “spherical harmonic” functions which are ubiquitous in physics (e. g., in the quantum-mechanical description of hydrogen-like atoms). This thesis is written with advanced undergraduate mathematics reader in mind. Part of the beauty of Lie groups is that it unites different areas of mathematics together. We will be using several topics from the standard undergraduate mathematics curriculum: linear, abstract algebra, analysis, metric topology, etc. Figuratively speaking, we may think of the study of Lie groups as the center of a wheel and each of the spokes as different branches of mathematics—all meeting together at the center of the wheel. The classical Lie groups preserve a bilinear form on a real or complex vector space, so we begin by defining bilinear forms and their associated matrix groups in chapter 2. Even if the reader is already familiar with the definitions, we recommend skimming through them to review the notation we will use throughout the rest of the thesis. In chapter 3, we define the orthogonal, unitary, and symplectic classical groups of matri- ces, which are perhaps familiar to the reader from linear algebra. For the purposes of this thesis, we call those the classical compact Lie groups. We also define general (abstract) Lie groups as differentiable manifolds with a group operation. Every Lie group has a Lie algebra attached to it, and these algebras will also play an important role in the thesis. It is possible to adopt a Lie algebraic approach to the study of the general aspects of representation theory of Lie groups; however, such approach would hide some (ultimately unavoidable) analytic and topological issues, as well as deny some of the benefits of a more unified approach. For these reasons, we eschew the study of representations of Lie algebras entirely. In chapter 4 we begin the study of some of the more elementary aspects of the represen- tation theory of compact groups. The normalized Haar integral plays a key role, by allowing CHAPTER 1. INTRODUCTION 3 to “average” over the group. Our primary interest for this thesis is to decompose represen- tations into irreducible subrepresentations. We next introduce the notion of characters of representations. Characters are very useful tools in understanding representation theory and we will make use of them often. In chapter 5 we study complex representations of connected abelian Lie groups (tori). Commutativity makes complex irreducible representations one-dimensional. This very im- portant special chapter of the representation theory of compact Lie groups is key to further study of the representations of non-abelian Lie groups. Chapter 6 revolves about the concept of maximal tori of a Lie group, that is, maximal connected abelian Lie subgroups. In a nutshell, restricting a representation of a compact connected Lie group to a maximal torus thereof does not, in principle, lose any information. Inchapter7, westudytheLiealgebras(anddualsthereof)ofthemaximaltoriofclassical compact Lie groups. The kernel of the covering map from the Lie algebra of a maximal torus to the torus is called the integer lattice; the latter is intimately related to the overall topology of the Lie group. The dual lattice (in the dual Lie algebra of the maximal torus) to the integer lattice is the weight lattice of the group; it plays a crucial role in the classification of irreducible representations. Inthefinalchapter8, welearnthatweights(elementsoftheweightlattice)ofaconnected compact Lie group correspond to irreducible representations. The main goal is to explain how this correspondence is established: the restriction of representations to a maximal torus is key. We also explain a relation between the operation of addition of two weights and the corresponding irreducible representations. We conclude the chapter and thesis with explicit examples that illustrate these correspondences. Chapter 2 Bilinear Forms 2.1 Bilinear and Sesquilinear Forms Definitions 2.1. Let V be a vector space over some field F. A bilinear form on V is a mapping B : V ×V → F; it takes two vectors and outputs a real scalar denoted by B(·,·). Bilinear forms satisfy the property that for all v,w ∈ V, and λ ∈ F, B(λv,w) = λB(v,w) and B(v +v ,w) = B(v ,w)+B(v ,w), 1 2 1 2 B(v,λw) = λB(v,w) and B(v,w +w ) = B(v,w )+B(v,w ). 1 2 1 2 For all v ∈ V and w ∈ V, a bilinear form is said to be symmetric if B(v,w) = B(w,v), skew-symmetric if B(v,w) = −B(w,v), and nondegenerate if B(v,·) = 0 implies v = 0. Nondegeneracy, in other words, means that if we have a nonzero vector v, then there exists some other vector w such that the bilinear form of v and w is nonzero. If we choose a basis B = {v ,v ,...v } of V, we may set up a matrix M that corresponds 1 2 n to the form given by M = [B] = [B(v ,v )]. B i j 4 CHAPTER 2. BILINEAR FORMS 5 Proposition 2.2. Let B be a bilinear form on a vector space V over F and M = [B(v ,v )] the matrix of the form with respect to a basis B. Let [v] and [w] be coordinate i j B B vectors of the vectors v and w respectively with basis B. Then B(v,w) = [v]TM[w] , B B where the superscript T denotes the matrix transpose. Definition 2.3. The standard inner product (cid:104)·,·(cid:105) of two vectors is an example of a bilinear form where the matrix of the form M with respect to the standard basis is the identity matrix I. Let v,w ∈ Rn, then B(v,w) = vTMw = vTIw = vTw = (cid:104)v,w(cid:105). Proposition 2.4. If B is a bilinear form on V and M = [B(v ,v )] is the matrix of the i j form, then i. B symmetric if and only if M = MT, ii. B skew-symmetric if and only if M = −MT, and iii. B nondegenerate if and only if M is nonsingular (invertible). Definitions 2.5. If B is a bilinear form on a vector space V, then B is either i. positive: B(v,v) (cid:62) 0, for all v ∈ V, ii. positive definite: B(v,v) > 0 for all v ∈ V such that v (cid:54)= 0, or iii. indefinite: B(v,v) > 0 for some v ∈ V, and B(w,w) < 0 for some w ∈ V. Definition 2.6. Let F be a field. The standard nondegenerate alternating bilinear form on F2n is a bilinear form whose matrix (with respect to the standard basis) is J = 2n   −I  n  , where I is the n×n identity matrix. The inner product between two vectors n   I n v and w for the symplectic form is (cid:104)v,w(cid:105) = vTJw.