Operators shrinking the Arveson spectrum

Similar documents
Fréchet differentiability of the norm of L p -spaces associated with arbitrary von Neumann algebras

CHAPTER X THE SPECTRAL THEOREM OF GELFAND

Peter Hochs. Strings JC, 11 June, C -algebras and K-theory. Peter Hochs. Introduction. C -algebras. Group. C -algebras.

The essential numerical range and the Olsen problem

Reflexivity and hyperreflexivity of bounded n-cocycle spaces and application to convolution operators. İstanbul Analysis Seminars, İstanbul, Turkey

TOEPLITZ OPERATORS. Toeplitz studied infinite matrices with NW-SE diagonals constant. f e C :

Banach Algebras where the Singular Elements are Removable Singularities

ON SPECTRA OF LÜDERS OPERATIONS

08a. Operators on Hilbert spaces. 1. Boundedness, continuity, operator norms

SPECTRAL THEOREM FOR SYMMETRIC OPERATORS WITH COMPACT RESOLVENT

MATH MEASURE THEORY AND FOURIER ANALYSIS. Contents

Invariant Subspaces and Complex Analysis

are Banach algebras. f(x)g(x) max Example 7.4. Similarly, A = L and A = l with the pointwise multiplication

Triple derivations on von Neumann algebras

Lecture III: Applications of Voiculescu s random matrix model to operator algebras

A Brief Introduction to Functional Analysis

Spectrum (functional analysis) - Wikipedia, the free encyclopedia

Preliminaries on von Neumann algebras and operator spaces. Magdalena Musat University of Copenhagen. Copenhagen, January 25, 2010

Gaussian automorphisms whose ergodic self-joinings are Gaussian

ON MATRIX VALUED SQUARE INTEGRABLE POSITIVE DEFINITE FUNCTIONS

Functional Analysis II held by Prof. Dr. Moritz Weber in summer 18

SPECTRAL THEORY EVAN JENKINS

Upper triangular forms for some classes of infinite dimensional operators

Real Analytic Version of Lévy s Theorem

Open Questions in Quantum Information Geometry

ELEMENTS OF PROBABILITY THEORY

On the uniform Opial property

MARCINKIEWICZ SPACES, COMMUTATORS AND NON-COMMUTATIVE GEOMETRY

Spectral Theory, with an Introduction to Operator Means. William L. Green

Math 123 Homework Assignment #2 Due Monday, April 21, 2008

Unbounded operators in the context of C -algebras

Spectral Measures, the Spectral Theorem, and Ergodic Theory

About Grupo the Mathematical de Investigación Foundation of Quantum Mechanics

CHAPTER VIII HILBERT SPACES

REFLEXIVITY OF THE SPACE OF MODULE HOMOMORPHISMS

THEOREMS, ETC., FOR MATH 515

Rational and H dilation

Spectral theory for linear operators on L 1 or C(K) spaces

Automorphic Equivalence Within Gapped Phases

ABSTRACT SCATTERING SYSTEMS: SOME SURPRISING CONNECTIONS

11. Spectral theory For operators on finite dimensional vectors spaces, we can often find a basis of eigenvectors (which we use to diagonalize the

Zero-one laws for functional calculus on operator semigroups

Spectral Mapping Theorem

Reminder Notes for the Course on Distribution Theory

Elliott s program and descriptive set theory I

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product

Global holomorphic functions in several non-commuting variables II

Topological vectorspaces

Spectral isometries into commutative Banach algebras

Introduction to Index Theory. Elmar Schrohe Institut für Analysis

MATH & MATH FUNCTIONS OF A REAL VARIABLE EXERCISES FALL 2015 & SPRING Scientia Imperii Decus et Tutamen 1

19. From a resolution of the identity to operators - and bac

Dualities in Mathematics: Locally compact abelian groups

Trace Class Operators and Lidskii s Theorem

Sung-Wook Park*, Hyuk Han**, and Se Won Park***

A STRUCTURED INVERSE SPECTRUM PROBLEM FOR INFINITE GRAPHS

TD 1: Hilbert Spaces and Applications

K theory of C algebras

On composition operators

ON THE SIMILARITY OF CENTERED OPERATORS TO CONTRACTIONS. Srdjan Petrović

Commutative Banach algebras 79

REPRESENTATION THEORY WEEK 7

Advanced functional analysis

ASSIGNMENT - 1, DEC M.Sc. (FINAL) SECOND YEAR DEGREE MATHEMATICS. Maximum : 20 MARKS Answer ALL questions. is also a topology on X.

Remarks on the Spectrum of Bounded and Normal Operators on Hilbert Spaces

MATH 113 SPRING 2015

On positive maps in quantum information.

C* ALGEBRAS AND THEIR REPRESENTATIONS

PRIME NON-COMMUTATIVE JB -ALGEBRAS

A non-commutative notion of separate continuity

Chapter 2 Unbounded Operators and Abstract Functional Calculus

10.1. The spectrum of an operator. Lemma If A < 1 then I A is invertible with bounded inverse

Operator Theory and Related Topics, Université de Lille 1, 2010 An application of generalized spectral operators

THE ALTERNATIVE DUNFORD-PETTIS PROPERTY FOR SUBSPACES OF THE COMPACT OPERATORS

Fredholmness of Some Toeplitz Operators

BERNARD RUSSO University of California, Irvine. Report on joint work with. Antonio M. Peralta Universidad de Granada, Spain

The Model Theory of C -algebras

Linear Algebra 1 Exam 2 Solutions 7/14/3

THE DAUGAVETIAN INDEX OF A BANACH SPACE 1. INTRODUCTION

The following definition is fundamental.

Abelian topological groups and (A/k) k. 1. Compact-discrete duality

Quantum Symmetric States

Borel complexity and automorphisms of C*-algebras

LINEAR PRESERVER PROBLEMS: generalized inverse

Functional Analysis I

Convergence Properties and Upper-Triangular Forms in Finite von Neumann Algebras

C -ALGEBRAS MATH SPRING 2015 PROBLEM SET #6

Problems in Linear Algebra and Representation Theory

THE BOUNDEDNESS BELOW OF 2 2 UPPER TRIANGULAR OPERATOR MATRICES. In Sung Hwang and Woo Young Lee 1

SEMICROSSED PRODUCTS OF THE DISK ALGEBRA

von Neumann algebras, II 1 factors, and their subfactors V.S. Sunder (IMSc, Chennai)

CHAPTER 1. Basic Analysis

Math Solutions to homework 5

arxiv: v3 [math-ph] 23 Sep 2014

Linear Algebra and Dirac Notation, Pt. 3

SPECTRAL RADIUS ALGEBRAS OF WEIGHTED CONDITIONAL EXPECTATION OPERATORS ON L 2 (F) Groups and Operators, August 2016, Chalmers University, Gothenburg.

MATH 241B FUNCTIONAL ANALYSIS - NOTES SPECTRAL THEOREM

0.2 Vector spaces. J.A.Beachy 1

ON GENERALIZED RIESZ POINTS. Robin Harte, Woo Young Lee and Lance L. Littlejohn

William Arveson Department of Mathematics University of California Berkeley CA 94720, USA. 29 February, 2000

Transcription:

Operators shrinking the Arveson spectrum A. R. Villena joint work with J. Alaminos and J. Extremera Departamento de Análisis Matemático Universidad de Granada BANACH ALGEBRAS 2009 STEFAN BANACH INTERNATIONAL MATHEMATICAL CENTER JULY 14-24, 2009 BEDLEWO, POLAND A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 1 / 31

The Arveson spectrum Let G be a locally compact abelian group, X be a complex Banach space, and τ : G inv ( B(X) ) be a bounded and strongly continuous group homomorphism. For every x X, the Arveson spectrum of x is defined as { } sp(x) = γ Ĝ : f (γ) = 0 for each f L 1 (G) with f (t)τ(t)xdt = 0. G A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 2 / 31

Examples Example (Support of the Fourier transform) Let G be a locally compact abelian group and let τ be the so-called regular representation of G on L 1 (G), τ : G B ( L 1 (G) ), [τ(t)f ](s) = f (t 1 s) (s, t G, f L 1 (G)). Then sp(f ) = supp f (f L 1 (G)). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 3 / 31

Example (Local spectrum of an operator) Let X be a complex Banach space. Given an operator T B(X), the local resolvent set ρ(t, x) of T at the point x X is defined as the union of all open subsets U of C for which there is an analytic function f : U X which satisfies (T z)f (z) = x for each z U. The local spectrum σ(t, x) of T at x is then defined as σ(t, x) = C \ ρ(t, x). 1 Let T B(X) be a doubly power bounded operator and consider Then sp(x) = σ(t, x) τ : Z B(X), τ(k) = T k (k Z). (x X). 2 Let T B(X) be a hermitian operator and consider the one-parameter group Then sp(x) = σ(t, x) τ : R B(X), τ(t) = exp(itt ) (t R). (x X). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 4 / 31

Example (Support of a spectral measure) Let A: D H H be a selfadjoint operator on a Hilbert space H. Then there is a projection-valued measure E on R such that A = + t de(t). For every x H, the map E( )x 2 defines a measure E x on the Borel subsets of R. Consider the one-parameter group τ : R U(H), τ(t) = exp(ita) (t R). Then sp(x) = supp ( E x ) (x H). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 5 / 31

The problem In many different contexts we find out operators Φ: X Y, where X and Y are complex Banach spaces, which shrink the Arveson espectrum in the sense that sp(φx) sp(x) (x X) for appropriate representations τ X and τ Y of a locally compact abelian group G on X and Y, respectively. The problem A typical example of operator Φ shrinking the spectrum is the one intertwining τ X and τ Y Φ τ X (t) = τ Y (t) Φ (t G). The standard problem consists in determining whether every operator shrinking the Arveson spectrum necessarily intertwines the representations. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 6 / 31

Our approach to the problem Suppose that Φ: X Y shrinks the Arveson spectrum. Pick t G and x X. We define a continuous bilinear map ϕ: A(T) A(T) Y, ϕ(f, g) = ( ( f (j)ĝ(k)τy (t j ) Φ τ X (t )x) ) k (f, g A(T)). j,k Z We check that sp(ϕ(f, g)) { } γ sp(x): γ(t) supp(f ) supp(g). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 7 / 31

Our approach to the problem: Disjointness vanishing bilinear maps on A(T) Accordingly, ϕ satisfies the property f, g A(T), supp(f ) supp(g) = }{{ ϕ(f, g) = 0. } disjointness vanishing The bilinear map ϕ induces a continuous linear operator Ψ: A(T 2 ) X, Ψ(f ) = ( ) f (j, k)ϕ(z j, z k ) f A(T 2 ) with the property that j,k Z f A(T 2 ), supp(f ) {(z, z): z T} = }{{} Ψ(f ) = 0. At this point the synthesis is coming into the scene! A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 8 / 31

The set is a set of synthesis for A(T 2 ). This means that every function f A(T 2 ) vanishing at can be approximated by a sequence (f n ) in A(T 2 ) with the property that f n vanishes on a neighborhood of. We then consider the function f A(T 2 ) defined by f (z, w) = z w (z, w T) and there exists a sequence (f n ) in A(T 2 ) with Accordingly, we have f = lim f n and supp(f n ) = (n N). 0 = lim Ψ(f n ) = Ψ(f ) = ϕ(z, 1) ϕ(1, z). }{{}}{{} ( ) ( ) τ Y (t) Φx Φ τ X (t)x A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 9 / 31

Our contribution Theorem Let G be a locally compact abelian group. Let X and Y be Banach spaces and let τ X : G B(X) and τ Y : G B(Y ) bounded and strongly continuous group homomorphisms. If Φ B(X, Y ) shrinks the Arveson spectrum, i.e. then Φ intertwines τ X and τ Y, i.e. sp(φx) sp(x) (x X), Φ τ X (t) = τ Y (t) Φ (t G). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 10 / 31

The secret of our success Theorem Let ϕ: A(T) A(T) X be a continuous bilinear map into some Banach space X with the property that f, g A(T), supp(f ) supp(g) = ϕ(f, g) = 0, then ϕ(z, 1) = ϕ(1, z). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 11 / 31

The Arveson spectrum for unbounded representations The theory of the Arveson spectrum still works for a non-quasianalytic representation τ : G B(X). This means that the weight function ω given by ω(t) = max{1, τ(t) } (t G) satisfies the Beurling-Domar condition + k= log ω(t k ) 1 + k 2 < (t G). For the definition of the Arveson spectrum of τ at a point x X we now replace the group algebra L 1 (G) by the weighted group algebra L 1 (G, ω). { } sp(x) = γ Ĝ : f (γ) = 0 for each f L 1 (G, ω) with f (t)τ(t)xdt = 0. G A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 12 / 31

Our theorem may fail in the case where the representation is unbounded. Example Consider the representations τ 1, τ 2 : Z B(C 2 ) given by Then and τ 1 (k) = ( ) k 1 1 and τ 0 1 2 (k) = I C2 (k Z). sp τ2 (x) sp τ1 (x) (x C 2 ) I C 2 τ 1 (k) τ 2 (k) I C 2 (k 0). In this case we have 1 + k 2 τ 1 (k) 2 + k 2 (k Z). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 13 / 31

For dealing with representations with polynomial growth we are required to involve disjointness vanishing bilinear maps on the weighted Fourier algebra { A α (T) = ϕ: A α (T) A α (T) X f C(T): f = k Z } f (k) (1 + k ) α < for some α 0. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 14 / 31

Theorem Let ϕ: A α (T) A α (T) X be a continuous bilinear map into some Banach space X with the property that f, g A α (T), supp(f ) supp(g) = ϕ(f, g) = 0. Then for each N > 2α. N i=0 ( ) N ( 1) i ϕ(z N i, z i ) = 0 i This time, ϕ gives rise to a continuous linear operator Ψ: A 2α (T 2 ) X. However, the diagonal may fail to be a set of synthesis for A 2α (T 2 ) because of the weight. We are thus required to take care of the function f at which we intend to apply Ψ. The so-called Beurling-Pollard type theorems assert, roughly speaking, that a function f admits synthesis if its growth is appropriate. We show that this is just the case for the function f (z, w) = (z w) N. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 15 / 31

An improvement to our contribution Theorem Let G be a locally compact abelian group. Let X and Y be Banach spaces and let τ X : G B(X) and τ Y : G B(Y ) strongly continuous group homomorphisms with polynomial growth, i.e. τy (t k ), τx (t k ) = O( k α ) as k for some α 0. If Φ B(X, Y ) is such that sp(φx) sp(x) (x X), then N i=0 whenever N > 2α. ( ) N ( 1) i τ Y (t N i ) Φ τ X (t i ) = 0 i (t G), A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 16 / 31

Applications: Translation invariant operators Corollary Let G be a locally compact abelian group. Then T : L 1 (G) L 1 (G) is translation invariant if and only if ) ) supp ( Tf supp( f (f L 1 (G)). Proof. We consider the regular representation of G on L 1 (G). Then ) ) supp ( Tf supp( f, } {{ } } {{ } sp(tf ) sp(f ) and therefore T shrinks the Arveson spectrum. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 17 / 31

Applications: Determining operators through the spectrum Corollary (Colojoară-Foiaş, 1968) Let X be a complex Banach space and S, T B(X) be invertible operators with polynomial growth. Suppose that Then σ(s, x) σ(t, x) x X. N i=0 ( ) N ( 1) i T N i S i = 0, i whenever N > 2α, where α is such that S k, T k = O( k α ) as k Consequently, 1 If S and T are doubly power bounded, then S = T. 2 If S and T commute, then (S T ) N = 0. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 18 / 31

Proof. We consider the representations τ S, τ T : Z B(X) given by τ S (k) = S k and τ T (k) = T k (k Z). Then σ(s, x) σ(t, x), }{{}}{{} sp τs (x) sp τt (x) which shows that the identity operator I X : X X shrinks the spectrum. Consequently, N i=0 ( ) N ( 1) i τ T (N i)i X τ S (i) i }{{} = 0 (t G), T N i S i A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 19 / 31

Corollary (I. Gelfand, 1941) Let T be a bounded linear operator on a complex Banach space X such that σ(t ) = {1}. If sup T k <, k Z then Example T = I X. It should be pointed out that Gelfand theorem fails in the case where the operator is not doubly power bounded. As a matter of fact, the operator T B(C 2 ) corresponding to the matrix ( ) 1 1 0 1 is different from the identity operator and nevertheless σ(t ) = {1}. In this case ( ) T k 1 k and 1 + k 0 1 2 T k 2 + k 2 k Z. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 20 / 31

Corollary (E. Hille, 1944) Let T be a bounded linear operator on a complex Banach space X such that σ(t ) = {1}. If T k = O( k n ) as k for some n 0, then for each integer N with N > n. (T I X ) N = 0. Example It should be pointed out that both Gelfand and Hille theorems fail for comparing the given operator with an operator different from the identity. Indeed, let S, T B(C 2 ) be the operators with matrices respectively. Then ( ) 1 1 0 1 and T S is far from being nilpotent. and ( ) 1 0, 1 1 σ(t ) = σ(s) = {1} A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 21 / 31

Theorem (Local version of Gelfand-Hille theorem) Let X be a complex Banach space and let T be an invertible operator on X such that T k = O( k α ) as k for some α 0. If x X is such that for some λ C, then σ(t, x) = {λ} (T λi X ) N x = 0 whenever N N is such that N > 2α. In particular, if T is doubly power bounded, then Tx = λx. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 22 / 31

Proof. First of all, it should be pointed out that λ σ(t, x) σ(t ) T. We define ϕ: A α (T) A α (T) X by ϕ(f, g) = g(λ) (k)t k Z f k x. Then ϕ is a disjointness vanishing continuous bilinear map. Our seminal result gives 0 = N i=0 ( ) N ( 1) i ϕ(z N i, z i ) = (T λi X ) N x. i }{{} λ i T N i x A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 23 / 31

Theorem (An improvement of the Colojoară-Foiaş theorem) Let X be a complex Banach space, S, T 1,..., T n, B(X) invertible operators with polynomial growth with the property that σ(s, x) n σ(t j, x) j=1 x X and that T 1,..., T n are pairwise commuting Then there exists N N such that C(S, T 1 ) N C(S, T n ) N I X = 0. Here the intertwiner C(S, T ) of S, T B(X) is defined by C(S, T ): B(X) B(X), C(S, T )A = SA AT A B(X). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 24 / 31

Applications: Local operators on function spaces and its non-commutative versions Theorem (A. C. Zaanen, 1975) Let Ω be a locally compact Hausdorff space and let T : C 0 (Ω) C 0 (Ω) be a bounded linear operator with the property that f C 0 (Ω), E Ω, f = 0 on E Tf = 0 on E. }{{} supp(tf ) supp(f ) Then there exists g C b (Ω) such that Tf = fg f C 0 (Ω). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 25 / 31

Theorem (Our non-commutative version) Let A be a (unital) C -algebra and let T : A A be a bounded linear operator with the property that x A, p projection in A, px = 0 }{{ ptx = 0. } left locality Then there exists a A such that Tx = xa x A. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 26 / 31

Proof. For every unitary u A, the map ϕ: A(T) A(T) A defined by ϕ(f, g) = f (j)ĝ(k)u j T (u k ) j,k Z is disjointness vanishing. This entails that T (u) = ut (1). Since A is generated by the unitaries, it follows that T (x) = x T (1) }{{} a x A. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 27 / 31

If we remove the requirement of A being unital, then the theorem still works though the element a defining the operator T lies in M(A), the multiplier algebra of A (i.e., the largest C -algebra containing A as an essential ideal). Of course, if A is unital, then M(A) = A. If A = C 0 (Ω), then M(A) = C b (Ω). If A = K(H), then M(A) = L(H). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 28 / 31

Theorem (A. C. Zaanen, 1975) Let (Ω, Σ, µ) be a σ-finite measure space, 1 p, and let T : L p (µ) L p (µ) be a bounded linear operator with the property that f L p (µ), E Σ, f = 0 a.e. on E }{{ Tf = 0 a.e. on E. } locality Then there exists g L (µ) such that Tf = fg f L p (µ). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 29 / 31

Theorem (Our noncommutative version) Let M be a von Neumann algebra with a (normal semifinite faithful) trace τ, 1 p, and let T : L p (M, τ) L p (M, τ) be a bounded linear operator with the property that x L p (M, τ), p projection in M, px = 0 ptx = 0. Then there exists a M such that Tx = xa x L p (M, τ). Let M = L (µ). Then integration with respect to µ gives a n.s.f. trace and L p (M) = L p (µ) (1 p ). Let M = L(H) and τ the usual trace on L(H). Then L p (M) is the Schatten class S p (H) (1 p < ). A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 30 / 31

Applications: Determining observables in quantum mechanics through the support 1 Every physical observable in quantum mechanics is mathematically represented by a selfadjoint operator. 2 The outcome of a measurement of an observable is of statistical nature. The value obtained in a measurement of the observable A on the state ψ is a random variable Âψ whose probability distribution is given by the spectral measure E ψ of A at ψ, so that the probability to obtain a value in R as the result of this measure is given by Corollary P [ Âψ ] = E( )ψ 2. Let A and B be observables on a quantum physical system with the property that ] ] P [Âψ = 1 P [ Bψ = 1. Then A = B. A. R. Villena (Granada) Operators shrinking the Arveson spectrum Banach Algebras 2009 31 / 31