Group Theory, Lattice Geometry, and Minkowski s Theorem

Similar documents
Equivalence Relations and Partitions, Normal Subgroups, Quotient Groups, and Homomorphisms

Kevin James. Quotient Groups and Homomorphisms: Definitions and Examp

A Crash Course in Topological Groups

CS 468: Computational Topology Group Theory Fall b c b a b a c b a c b c c b a

Lecture 7.3: Ring homomorphisms

Metric spaces and metrizability

A Little Beyond: Linear Algebra

Lecture Note of Week 2

Modern Algebra Prof. Manindra Agrawal Department of Computer Science and Engineering Indian Institute of Technology, Kanpur

BRAID GROUPS ALLEN YUAN. 1. Introduction. groups. Furthermore, the study of these braid groups is also both important to mathematics

Chapter 2 Linear Transformations

AM 106/206: Applied Algebra Madhu Sudan 1. Lecture Notes 11

0.2 Vector spaces. J.A.Beachy 1

Math 581 Problem Set 8 Solutions

1 Differentiable manifolds and smooth maps

Lecture 10: A (Brief) Introduction to Group Theory (See Chapter 3.13 in Boas, 3rd Edition)

BASIC GROUP THEORY : G G G,

2MA105 Algebraic Structures I

GEOMETRIC CONSTRUCTIONS AND ALGEBRAIC FIELD EXTENSIONS

Abstract Vector Spaces and Concrete Examples

THE REPRESENTATION THEORY, GEOMETRY, AND COMBINATORICS OF BRANCHED COVERS

Introduction to Group Theory

ROTATIONS, ROTATION PATHS, AND QUANTUM SPIN

Chapter 3. Introducing Groups

ENTRY GROUP THEORY. [ENTRY GROUP THEORY] Authors: started Mark Lezama: October 2003 Literature: Algebra by Michael Artin, Mathworld.

The Hurewicz Theorem

The First Isomorphism Theorem

MATH 436 Notes: Homomorphisms.

Algebraic Topology M3P solutions 2

Partial, Total, and Lattice Orders in Group Theory

Fall /29/18 Time Limit: 75 Minutes

2. normal subgroup and quotient group We begin by stating a couple of elementary lemmas Lemma. Let A and B be sets and f : A B be an onto

Author: Bob Howlett Group representation theory Lecture 1, 28/7/97. Introduction

Group theory applied to crystallography

Locally definable groups and lattices

Isomorphisms and Well-definedness

Part IV. Rings and Fields

Computational Approaches to Finding Irreducible Representations

THE FUNDAMENTAL GROUP AND SEIFERT-VAN KAMPEN S THEOREM

0. Introduction 1 0. INTRODUCTION

MATH 436 Notes: Cyclic groups and Invariant Subgroups.

Lecture 2: Groups. Rajat Mittal. IIT Kanpur

book 2005/1/23 20:41 page 132 #146

Lemma 1.3. The element [X, X] is nonzero.

AN AXIOMATIC FORMATION THAT IS NOT A VARIETY

10. Smooth Varieties. 82 Andreas Gathmann

sset(x, Y ) n = sset(x [n], Y ).

Mathematics 331 Solutions to Some Review Problems for Exam a = c = 3 2 1

Introduction to Arithmetic Geometry Fall 2013 Lecture #23 11/26/2013

Cambridge University Press The Mathematics of Signal Processing Steven B. Damelin and Willard Miller Excerpt More information

Algebraic Varieties. Notes by Mateusz Micha lek for the lecture on April 17, 2018, in the IMPRS Ringvorlesung Introduction to Nonlinear Algebra

DS-GA 1002 Lecture notes 0 Fall Linear Algebra. These notes provide a review of basic concepts in linear algebra.

Section 15 Factor-group computation and simple groups

3. The Sheaf of Regular Functions

Cosets, factor groups, direct products, homomorphisms, isomorphisms

HOMOLOGY AND COHOMOLOGY. 1. Introduction

Lectures - XXIII and XXIV Coproducts and Pushouts

GROUP ACTIONS RYAN C. SPIELER

Scott Taylor 1. EQUIVALENCE RELATIONS. Definition 1.1. Let A be a set. An equivalence relation on A is a relation such that:

Some algebraic properties of. compact topological groups

Solution Outlines for Chapter 6

1. Introduction to commutative rings and fields

HOW DO ULTRAFILTERS ACT ON THEORIES? THE CUT SPECTRUM AND TREETOPS

Abstract Algebra Study Sheet

Section III.15. Factor-Group Computations and Simple Groups

Math 676. A compactness theorem for the idele group. and by the product formula it lies in the kernel (A K )1 of the continuous idelic norm

From Wikipedia, the free encyclopedia

6 More on simple groups Lecture 20: Group actions and simplicity Lecture 21: Simplicity of some group actions...

HOW TO LOOK AT MINKOWSKI S THEOREM

Affine Connections: Part 2

Summer Algebraic Geometry Seminar

Corrections to Introduction to Topological Manifolds (First edition) by John M. Lee December 7, 2015

AN INTRODUCTION TO AFFINE SCHEMES

CONSEQUENCES OF THE SYLOW THEOREMS

Algebraic Number Theory

Algebraic Topology I Homework Spring 2014

MATH 25 CLASS 21 NOTES, NOV Contents. 2. Subgroups 2 3. Isomorphisms 4

COMBINATORIAL GROUP THEORY NOTES

COMPLEX ANALYSIS Spring 2014

Extra exercises for algebra

Commutative Banach algebras 79

Math 594. Solutions 5

THE FUNDAMENTAL GROUP AND CW COMPLEXES

HYPERBOLIC DYNAMICAL SYSTEMS AND THE NONCOMMUTATIVE INTEGRATION THEORY OF CONNES ABSTRACT

Introduction to Groups

Recall: Properties of Homomorphisms

1.1 The program of Lie

A Generalization of Wilson s Theorem

Valuation Rings. Rachel Chaiser. Department of Mathematics and Computer Science University of Puget Sound

Homomorphisms. The kernel of the homomorphism ϕ:g G, denoted Ker(ϕ), is the set of elements in G that are mapped to the identity in G.

Outline. Linear maps. 1 Vector addition is commutative: summands can be swapped: 2 addition is associative: grouping of summands is irrelevant:

NOTES ON FINITE FIELDS

Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra

Vectors. January 13, 2013

Math 210C. The representation ring

AN INTRODUCTION TO THE FUNDAMENTAL GROUP

LECTURE 3: RELATIVE SINGULAR HOMOLOGY

Rings. Chapter 1. Definition 1.2. A commutative ring R is a ring in which multiplication is commutative. That is, ab = ba for all a, b R.

Solutions to Problem Set 1

ALGEBRAIC GROUPS J. WARNER

Transcription:

Group Theory, Lattice Geometry, and Minkowski s Theorem Jason Payne Physics and mathematics have always been inextricably interwoven- one s development and progress often hinges upon the other s. It is that which motivates this primer on group theory and lattice geometry in the hopes of presenting the mathematical tools refined by explorations into a field known as algebraic number theory in a context more easily accessible to a physicist studying lattice structure. The focus will be on clarity of exposition and a grounding in examples of the more elusive concepts. This paper is separated into two parts. First, we will introduce a selection of the vital foundational concepts in the theory of groups. Following this will be a basic introduction to lattice geometry: bringing to light some crucial concepts like that of lattice vectors, the interplay between (n-dimensional) lattices and the (n-dimensional) Torus T n, fundamental domains, and convexity. This will culminate in the proof of Minkowski s Theorem which states that if a bounded, symmetric, convex subset X of R n satisfies the following condition on its volume, V(X) > 2 n V(D), where D is the fundamental domain of an n-dimensional lattice L, then X must contain a non-zero lattice vector. This theorem has far reaching implications in the algebraic theory of numbers, as well as their geometry, and consequently in the lattice-entrenched theory of condensed matter physics. It is the hope of the author that with this brief introduction the reader can involve themselves in a deeper study of the theory of solids relying upon a rigorous mathematical formalism of the internal geometry of crystals.

2 I. INTRODUCTION TO GROUP THEORY We begin with a very brief review of the essential elements of group theory. Although group theory is, by itself, a substantial and vital portion of modern mathematics, we will focus on the foundational concepts relevant to the discussion that follows. In the interest of clarity, the more subtle details (for example, the potential existence of a left-inverse that is not a right-inverse) will be neglected. Of course, all of this would be for naught without first introducing the notion of a group. Definition 1. Given a (possibly infinite) collection of objects G = {g 1,,g n, } coupled with an operation : G G G, we call G a group if it satisfies the following: (i) (Closure) Given any two elements g,h G, we have that g h G. (ii) (Group Identity) There exists an element e G so that e is called the identity element of G. (iii) (Inverses) Given any g G there exists an element h G so that g e = e g = g. ( g G) g h = h g = e. h is called the (two-sided) inverse of g and is denoted g 1. This is merely an abstraction of an idea with which we are all familiar. The focus of group theory is on examining precisely how the general principles of arithmetic or multiplication apply in other situations it is an analysis of the blueprints of arithmetic and multiplicative structures. It should be pointed out that often times care is necessary in describing a group, as the same set of elements can have multiple, distinct operations defined on it. A few examples that should be familiar to the reader are given below. Ex 1. Consider R with the usual notion of addition (or, more generally, R n with the usual notion of vector addition). The (additive) identity element is given by e = 0, and the (additive) inverse of a real number a is simply a. Similarly, one could use the real numbers with multiplication. This latter remark highlights what was said above it is, in some cases, important to signify the operation one is interested in. In the context of this example, one can achieve this by, rather than simply referring to the group R, talking about the group (R,+) or the group (R, ), where the first term in the pair identifies the set of interest, and the second denotes the operation. Ex 2. The physical Hilbert space associated with a quantum mechanical system (or, more generally, any vector space) has contained within it a group structure that of the states, given by kets, and addition operation to generate (potentially) new states. Indeed, recall that in the definition of the underlying vector space of a Hilbert space the closure, identity, and inverse properties are included. Ex 3. Consider S = {z C z = 1} with : S S S defined as the usual complex multiplication: (a+ib) (c+id) = (ac bd)+i(ad+bc). This is called the circle group and it is related to rotations. Definition 2. Let (G, ) and (H,+) be groups. Then a map f : G H is called a (group) homomorphism if for all g 1,g 2 G we have f(g 1 g 2 ) = f(g 1 )+f(g 2 ). If, in addition, f is one-to-one, onto, or both then it is called a monomorphism, epimorphism, and isomorphism, respectively. One point of clarification is that the above operations and + do not necessarily have any relation to the usual notions of addition and multiplication they simply signify that we (potentially) have two different operations in play here. So what a homomorphism really is, is a map which, in a sense, preserves the operations of the group it is a map for which you can either operate on the arguments in G before applying the map, or apply the map to each of the arguments separately and then operate in H on the results.

Definition 3. Let H G. Then H is called a subgroup of G if H satisfies (i) (iii) above when considered as a set with the same operation as G. That is to say that it is closed under the operation of G, it includes the identity element of G, and h H implies that h 1 H, where h 1 is the same inverse as when h is considered as an element of G. An equivalent, and typically useful, characterization of the idea of a subgroup is that it is a subset H for which, given any h 1,h 2 H we also have that h 1 h 1 2 H. We now introduce two very special subgroups related to any given homomorphism. Definition 4. Let f : G H be a group homomorphism. Then the subsets and kerf = {g G f(g) = e H } imf = {h H f(g) = h for some g G} are called the kernel of f and the image of f, respectively. Theorem 1. Let N < G; then the following conditions on N are equivalent: (i) gn = Ng, g G; (ii) gng 1 N, g G; (iii) gng 1 = N, g G. A subgroup N which satisfies any of the above conditions is called a normal subgroup, and this is denoted by N G. Proof. 3 (i) (ii): Suppose that gn = Ng for all g G. This means that for a fixed g G and n N, there exists an element n N so that Multiplying on the right by g 1 we have that gn = n g. gng 1 = n gg 1 = n N. Repeating this for each n N and g G yields gng 1 N, as desired. (ii) (iii): Suppose that gng 1 N. To demonstrate equality we need to establish the reverse containment, i.e. show N gng 1. The key to this is to realize that (ii) applies to g 1 as well, which allows us to write, for any n N thus N gng 1 as well, so we have N = gng 1. n = (gg 1 )n(gg 1 ) = g(g 1 ng)g 1 = gn g 1 gng 1 ; (iii) (i): This is immediate just multiply both sides of (iii) on the right by g: (gn g 1 )g = Ng gn = Ng. The above implications give us a closed loop, of sorts, of implications which establishes the equivalence of the three claims: (i) (ii) (iii)

Having provided a few equivalent ways of thinking about normal subgroups, we are now able to introduce another important concept in group theory that of the quotient group. Definition 5. Given a normal subgroup N G, then we can define a new group, G/N, associated with N as the set of all (left) cosets of N in G, that is The operation which turns this into a group is given by G/N = {gn g G}. (gn) (hn) = (gh)n. (g,h G) An important technical detail to keep in mind is that in order for the set G/N to be a group, N must be a normal subgroup. For an intuitive understanding of what the quotient group is, one can think of it as an identification you are, in some sense, taking all the elements of the subgroup N and identifying them with each other. You treat them as if they are the same element in order to focus on the action N has with other elements of the group G. Next, we point out that some of the most important theoretic tools group-theorists have at their disposal at that of the so-called isomorphism theorems. We will only make use of the first one: Theorem 2. Let ϕ : G H be a homomorphism of groups. Then (i) kerϕ G; (ii) im(ϕ) < H; (iii) (First Isomorphism Theorem) im(ϕ) = G/ker(ϕ). Proof. 4 (i) Let x kerϕ and g G. By the definition of homomorphism we have ϕ(gxg 1 ) = ϕ(g)ϕ(x)ϕ(g 1 ) = ϕ(g)eϕ(g) 1 = e; thus gxg 1 kerϕ. Since x was an arbitrary element of the kernel this implies that g(kerϕ)g 1 kerϕ. Therefore by Theorem 1(ii) we have that kerϕ G, as desired. (ii) Let a,b imϕ. Then there exist some g,h G so that ϕ(g) = a and ϕ(h) = b. However, this means that hence imϕ < H. ab 1 = ϕ(g)ϕ(h) 1 = ϕ(g)ϕ(h 1 ) = ϕ(gh 1 ) imϕ; (iii) (Sketch) The situation we are dealing with can be codified by the following diagram: π G G/kerϕ ϕ ϕ imϕ where π : G G/kerϕ is the quotient map and ϕ is the isomorphism we are looking for. The idea is to define ϕ so that the above diagram is commutative, i.e. one can travel around the diagram along either of the two paths (ϕ or ϕ π) with the same result. Therefore, we let ϕ : G/kerϕ imϕ be defined as ϕ(g(kerϕ)) = ϕ(g). Since this proof is included simply to illuminate a common approach in abstract algebraic problems through the above definition, we will, for the sake of brevity, leave the remainder to the reader. The interested reader is invited to explore the rich framework that group-theoretic investigations has illuminated within modern mathematics. A particularly comprehensive and sophisticated treatise on the subject is given by [3]; alternatively, an intuitive, example-driven discussion can be found in [1].

5 II. INTRODUCTION TO LATTICE GEOMETRY [Note: This section will closely follow [6]. Additional insight came from [4] and [5], as well as the author s understanding of the subject based on a course in algebraic number theory using the text [2].] The starting point for this section is the definition of a lattice a concept closely related to, and occasionally analogous to, that of a vector space. For simplicity, we will consider only lattice subsets of R n. Definition 6. Let {e 1,,e n } be a set of linearly independent vectors in R n. Then the additive subgroup L < R n generated by e 1,,e n is called an n-dimensional lattice. This subgroup consists of all the elements v R n which can be written in the form v = n n k e k. (n k Z) k=1 Associated with any lattice is a particular region in R n, called the fundamental domain (we will denote this region by D), which consists of all elements w R n of the form w = n a k e k. (0 a k < 1) k=1 This is what is known as the unit cell in the theory of solids. We can go immediately from this definition to the following theorem: Theorem 3. Let L be an n-dimensional lattice in R n. The L is isomorphic to the n-dimensional torus T n. Sketch. We begin by viewing the generating vectors of the lattice e 1,,e n as a basis for R n. This means, in particular, that any vector v R n can be represented as We can define a map φ : R n T n by v = n a k e k. (a k R) k=1 φ(v) = (e 2πia1,e 2πia2,,e 2πian ). Then we have immediately that φ is a homomorphism, and is onto. There are two important details to notice the first is that if v L then a k Z for all k, so φ(v) = (0,0,,0), i.e. we have that kerφ = L. Secondly, since φ is onto we have that imφ = T n ; hence, by the First Isomorphism Theorem for groups, we have an isomorphism as desired. R n /L = R n /kerφ = imφ = T n, This theorem may, on the face of it, appear to be more of a mathematical curiosity than a concept deeply rooted in the study of lattices; however, it turns out, this theorem provides the connection between lattices and the second topic of this section geometry. One of the foundational objects of interest in any geometrical field of mathematics is that of measure how one can assign appropriate notions of length, area, volume, and their n-dimensional counterparts to the elements of the topological space of interest. Fortunately, in the case of lattice geometry we have just established a connection between the abstract concept of an additive subgroup of R n and the more concrete (and well understood) object known as a torus. It is through this correspondence that we can develop a suitable perspective on the meaning of (n-dimensional) volume in a lattice, as seen in the following definition.

Definition 7. Recall that the volume of a (measurable) subset X R n is given by integrating against the Lebesgue measure (which for simple, well-behaved subsets like the torus is equivalent to the more familiar Riemann integral of the volume element dx 1 dx n ): v(x) = dx 1 dx n. X To extend this to the more indefinite idea of a lattice, we utilize the above correspondence in order to pullback the notion of volume through φ. Thus, the volume of a subset X D, where D is the fundamental domain of the lattice L, is defined to be v(x) = v(φ 1 (X)) = dx 1 dx n. φ 1 (X) The following theorem (stated without proof) provides a useful characterization of the relationship between the volume of a subset X R n and that of a subset of the lattice L. Theorem 4. Let φ : R n T n be defined as in Theorem 3. If X is a bounded subset of R n, v(x) exists, and v(φ(x)) v(x), then φ X : X T n is not one-to-one. Finally, before moving onto the primary result presented in this paper, we will define a few more concepts common to the study of geometry. Definition 8. A subset X R n is said to be convex if x,y X implies that λx+(1 λ)y X for all λ [0,1]. To aid in the intuitive understanding of convexity, note that this simply means the line segment connecting x to y also lies entirely within X. X is said to be symmetric if x X implies that x X. This simply means that X is invariant under reflection through the origin. Furthermore, X is called bounded if there exists an n-dimensional sphere S (of finite radius) so that X S. Since convexity is, in our experience, the most difficult to grasp, Fig. 1 provides a few illustrations to aid in understanding it. 6 FIG. 1. Examples of a Convex and a Non-Convex Set x y Convex Non-convex Theorem 5 (Minkowski, 1896). Let D be the fundamental domain of an n-dimensional lattice L. Further, let X be a bounded, symmetric, convex subset of L. If V(X) > 2 n V(D) then X contains at least one non-zero lattice vector (i.e. element of L). Proof. We begin by choosing a basis of L, say v 1,,v n and consider the n-dimensional lattice 2L generated by 2v 1,,2v n and its fundamental domain 2D. We can see that v(2d) = 2 n v(d) since we have simply doubled each of the dimensions, resulting in an overall change of 2 n one factor of 2 for each dimension (just think of simple case of the volume of a unit sphere and what happens when you double the length of each side). Notice, however, that this doubling does not affect the isomorphism with the n-dimensional torus, i.e. we have R n /2L = T n,

7 which, when coupled with the above discussion of the volume of 2D, yields But this means that v(t n ) = v(2d) = 2 n v(d). v(φ(x)) v(t n ) = 2 n v(d) < v(x), so by Theorem 4 we have that φ X is not one-to-one. This furnishes some distinct x 1,x 2 X so that φ(x 1 ) = φ(x 2 ), which is equivalent to x 1 x 2 2L. Another way of thinking of this condition is to notice that Thus convexity and symmetry of X ensure that 1 2 (x 1 x 2 ) = 1 2 x 1 + 1 2 (x 1 x 2 ) L. ( 1 1 ) ( x 2 ) X. 2 This mean that 1 2 (x 1 x 2 ) is a lattice vector in X which is nonzero since x 1 and x 2 are distinct; thus we have found the element we were looking for. This result may seem (much like the theorem on the isomorphism between the quotient group R n /L and the torus T n ) to be inconsequential. It is this theorem, however, upon which nearly all aspects of lattice geometry lie, as well as a vast wealth of results in algebraic number theory. If one wishes to get a grasp on the geometry of the internal structure of crystals, then understanding this theorem provides a crucial stepping stone in that direction. [1] D.S. Dummit and R.M. Foote. Abstract Algebra. Wiley, 3rd edition, 2003. [2] A. Frölich and M.J. Taylor. Algebraic Number Theory. Cambridge Studies in Advanced Mathematics, 27. Cambridge University Press, 1993. [3] T. Hungerford. Algebra. Graduate Texts in Mathematics, 73. Springer, 1980. [4] H.W. Lenstra. Lattices, 2008. http://www.math.leidenuniv.nl/~psh/antproc/06hwl.pdf. [5] J.S. Milne. Algebraic number theory (v3.03). www.jmilne.org/math/, 2011. [6] R. Virk. The Geometry of Numbers. http://www.math.ucdavis.edu/~virk/notes/pre08/pdf/geometryofnumbers.pdf.