On m-sectorial Schrödinger-type operators with singular potentials on manifolds of bounded geometry

Similar documents
SELF-ADJOINTNESS OF SCHRÖDINGER-TYPE OPERATORS WITH SINGULAR POTENTIALS ON MANIFOLDS OF BOUNDED GEOMETRY

Title: Localized self-adjointness of Schrödinger-type operators on Riemannian manifolds. Proposed running head: Schrödinger-type operators on

THE FORM SUM AND THE FRIEDRICHS EXTENSION OF SCHRÖDINGER-TYPE OPERATORS ON RIEMANNIAN MANIFOLDS

On m-accretive Schrödinger operators in L p -spaces on manifolds of bounded geometry

A PROPERTY OF SOBOLEV SPACES ON COMPLETE RIEMANNIAN MANIFOLDS

Sectorial Forms and m-sectorial Operators

Generalized Schrödinger semigroups on infinite weighted graphs

THE ESSENTIAL SELF-ADJOINTNESS OF SCHRÖDINGER OPERATORS WITH OSCILLATING POTENTIALS. Adam D. Ward. In Memory of Boris Pavlov ( )

Wave equation on manifolds and finite speed of propagation

Periodic Schrödinger operators with δ -potentials

ON THE SPECTRUM OF THE DIRICHLET LAPLACIAN IN A NARROW STRIP

Wentzell Boundary Conditions in the Nonsymmetric Case

Heat kernels of some Schrödinger operators

QUASI-UNIFORMLY POSITIVE OPERATORS IN KREIN SPACE. Denitizable operators in Krein spaces have spectral properties similar to those

The Kato square root problem on vector bundles with generalised bounded geometry

The spectral zeta function

Rough metrics, the Kato square root problem, and the continuity of a flow tangent to the Ricci flow

The Kato square root problem on vector bundles with generalised bounded geometry

THE L 2 -HODGE THEORY AND REPRESENTATION ON R n

SELF-ADJOINTNESS OF DIRAC OPERATORS VIA HARDY-DIRAC INEQUALITIES

REMARKS ON THE MEMBRANE AND BUCKLING EIGENVALUES FOR PLANAR DOMAINS. Leonid Friedlander

MULTIPLE SOLUTIONS FOR AN INDEFINITE KIRCHHOFF-TYPE EQUATION WITH SIGN-CHANGING POTENTIAL

POINTWISE BOUNDS ON QUASIMODES OF SEMICLASSICAL SCHRÖDINGER OPERATORS IN DIMENSION TWO

PERTURBATION THEORY FOR NONLINEAR DIRICHLET PROBLEMS

ABSOLUTELY CONTINUOUS SPECTRUM OF A TYPICAL SCHRÖDINGER OPERATOR WITH A SLOWLY DECAYING POTENTIAL

L -uniqueness of Schrödinger operators on a Riemannian manifold

SUBELLIPTIC CORDES ESTIMATES

GORDON TYPE THEOREM FOR MEASURE PERTURBATION

Radial Symmetry of Minimizers for Some Translation and Rotation Invariant Functionals

SOLVABILITY RELATIONS FOR SOME NON FREDHOLM OPERATORS

On the Virial Theorem in Quantum Mechanics

EXISTENCE OF SOLUTIONS FOR KIRCHHOFF TYPE EQUATIONS WITH UNBOUNDED POTENTIAL. 1. Introduction In this article, we consider the Kirchhoff type problem

On the Solvability Conditions for a Linearized Cahn-Hilliard Equation

Inverse scattering problem with underdetermined data.

Isoperimetric Inequalities for the Cauchy-Dirichlet Heat Operator

HOMEOMORPHISMS OF BOUNDED VARIATION

An Example of Embedded Singular Continuous Spectrum for One-Dimensional Schrödinger Operators

EXPONENTIAL ESTIMATES FOR QUANTUM GRAPHS

arxiv: v1 [math.ap] 5 Nov 2018

On projective classification of plane curves

BASIC MATRIX PERTURBATION THEORY

arxiv:math/ v1 [math.ap] 22 May 2006

Spectral Gap and Concentration for Some Spherically Symmetric Probability Measures

On Schrödinger equations with inverse-square singular potentials

On a generalized Sturm theorem

ATTRACTORS FOR SEMILINEAR PARABOLIC PROBLEMS WITH DIRICHLET BOUNDARY CONDITIONS IN VARYING DOMAINS. Emerson A. M. de Abreu Alexandre N.

APPROXIMATION OF MOORE-PENROSE INVERSE OF A CLOSED OPERATOR BY A SEQUENCE OF FINITE RANK OUTER INVERSES

On some weighted fractional porous media equations

Mem. Differential Equations Math. Phys. 36 (2005), T. Kiguradze

A LIOUVILLE THEOREM FOR p-harmonic

The effect of Group Velocity in the numerical analysis of control problems for the wave equation

ESTIMATES OF DERIVATIVES OF THE HEAT KERNEL ON A COMPACT RIEMANNIAN MANIFOLD

Applied Math Qualifying Exam 11 October Instructions: Work 2 out of 3 problems in each of the 3 parts for a total of 6 problems.

Stability of the Magnetic Schrödinger Operator in a Waveguide

SMOOTHING ESTIMATES OF THE RADIAL SCHRÖDINGER PROPAGATOR IN DIMENSIONS n 2

Proceedings of the 5th International Conference on Inverse Problems in Engineering: Theory and Practice, Cambridge, UK, 11-15th July 2005

The Mollifier Theorem

NOTES ON NEWLANDER-NIRENBERG THEOREM XU WANG

Edge-cone Einstein metrics and Yamabe metrics

Traces for star products on symplectic manifolds

A SHARP STABILITY ESTIMATE IN TENSOR TOMOGRAPHY

1.4 The Jacobian of a map

SUPERCONVERGENCE PROPERTIES FOR OPTIMAL CONTROL PROBLEMS DISCRETIZED BY PIECEWISE LINEAR AND DISCONTINUOUS FUNCTIONS

A Walking Tour of Microlocal Analysis

Universität Regensburg Mathematik

EXISTENCE OF SOLUTIONS TO ASYMPTOTICALLY PERIODIC SCHRÖDINGER EQUATIONS

Reproducing formulas associated with symbols

ASYMPTOTIC STRUCTURE FOR SOLUTIONS OF THE NAVIER STOKES EQUATIONS. Tian Ma. Shouhong Wang

A collocation method for solving some integral equations in distributions

A nodal solution of the scalar field equation at the second minimax level

Analysis in weighted spaces : preliminary version

Trotter s product formula for projections

MULTI-VALUED BOUNDARY VALUE PROBLEMS INVOLVING LERAY-LIONS OPERATORS AND DISCONTINUOUS NONLINEARITIES

A note on W 1,p estimates for quasilinear parabolic equations

THE FUNDAMENTAL THEOREM OF ALGEBRA VIA PROPER MAPS

ALEKSANDROV S THEOREM: CLOSED SURFACES WITH CONSTANT MEAN CURVATURE

MULTIPLE POSITIVE SOLUTIONS FOR KIRCHHOFF PROBLEMS WITH SIGN-CHANGING POTENTIAL

Determinant of the Schrödinger Operator on a Metric Graph

SCHRÖDINGER OPERATORS WITH PURELY DISCRETE SPECTRUM

satisfying the following condition: If T : V V is any linear map, then µ(x 1,,X n )= det T µ(x 1,,X n ).

Spectral Theory of Orthogonal Polynomials

GENERATORS WITH INTERIOR DEGENERACY ON SPACES OF L 2 TYPE

Transparent connections

Global regularity of a modified Navier-Stokes equation

Traces, extensions and co-normal derivatives for elliptic systems on Lipschitz domains

THE TWO-PHASE MEMBRANE PROBLEM REGULARITY OF THE FREE BOUNDARIES IN HIGHER DIMENSIONS. 1. Introduction

arxiv: v2 [math.fa] 17 May 2016

Lecture 4: Harmonic forms

23 Elements of analytic ODE theory. Bessel s functions

Operators with numerical range in a closed halfplane

THE BEST CONSTANT OF THE MOSER-TRUDINGER INEQUALITY ON S 2

HIGHER INTEGRABILITY WITH WEIGHTS

ON BOUNDEDNESS OF MAXIMAL FUNCTIONS IN SOBOLEV SPACES

ON THE SPECTRUM OF NARROW NEUMANN WAVEGUIDE WITH PERIODICALLY DISTRIBUTED δ TRAPS

Applications of Mathematics

here, this space is in fact infinite-dimensional, so t σ ess. Exercise Let T B(H) be a self-adjoint operator on an infinitedimensional

Algebraic Sum of Unbounded Normal Operators and the Square Root Problem of Kato.

A Class of Shock Wave Solutions of the Periodic Degasperis-Procesi Equation

A symmetry-based method for constructing nonlocally related partial differential equation systems

Cohomology of the Mumford Quotient

Transcription:

Comment.Math.Univ.Carolin. 45,1(2004)91 100 91 On m-sectorial Schrödinger-type operators with singular potentials on manifolds of bounded geometry Ognjen Milatovic Abstract. WeconsideraSchrödinger-typedifferentialexpression H V = +V,where isac -boundedhermitianconnectiononahermitianvectorbundle Eofbounded geometryoveramanifoldofboundedgeometry(m, g)withmetric gandpositive C - bounded measure dµ, and V is a locally integrable section of the bundle of endomorphismsof E. Wegiveasufficientconditionfor m-sectorialityofarealizationof H V in L 2 (E). IntheproofweusegeneralizedKato sinequalityaswellasaresultonthe positivityof u L 2 (M)satisfyingtheequation( M + b)u=ν,where M isthescalar Laplacianon M, b >0isaconstantand ν 0isapositivedistributionon M. Keywords: Schrödinger operator, m-sectorial, manifold, bounded geometry, singular potential Classification: Primary 35P05, 58J50; Secondary 47B25, 81Q10 1. Introduction and the main result 1.1 The setting. Let(M, g)beac Riemannianmanifoldwithoutboundary,withmetric ganddimm = n. Wewillassumethat M isconnected. We will also assume that M has bounded geometry. Moreover, we will assume that wearegivenapositive C -boundedmeasure dµ,i.e.inanylocalcoordinates x 1, x 2,...,x n thereexistsastrictlypositive C -boundeddensity ρ(x)suchthat dµ=ρ(x)dx 1 dx 2... dx n. Let EbeaHermitianvectorbundleover M.Wewillassumethat Eisabundle of bounded geometry(i.e. it is supplied by an additional structure: trivializations of E on every canonical coordinate neighborhood U such that the corresponding matrixtransitionfunctions h U,U onallintersections U U ofsuchneighborhoodsare C -bounded,i.e.allderivatives y αh U,U (y),where αisamultiindex, withrespecttocanonicalcoordinatesareboundedwithbounds C α whichdonot dependonthechosenpair U, U ). Wedenoteby L 2 (E)theHilbertspaceofsquareintegrablesectionsof Ewith respect to the scalar product (1.1) (u, v) = u(x), v(x) dµ(x). M Here, denotesthefiberwiseinnerproductin E x.

92 O. Milatovic Inwhatfollows, C (E)denotessmoothsectionsof E,and C c (E)denotes smooth compactly supported sections of E. Let : C (E) C (T M E) beahermitianconnectionon E whichis C -boundedasalineardifferential operator, i.e. in any canonical coordinate system U(with the chosen trivializations of E U and(t M E) U ), iswrittenintheform = a α (y) y α, α 1 where αisamultiindex,andthecoefficients a α (y)arematrixfunctionswhose derivatives β y a α (y)foranymultiindex βareboundedbyaconstant C β which does not depend on the chosen canonical neighborhood. We will consider a Schrödinger type differential expression of the form H V = +V, where V isameasurablesectionofthebundleendeofendomorphismsof E. Here : C (T M E) C (E) isadifferentialoperatorwhichisformallyadjointto withrespecttothescalar product(1.1). Ifwetake =d,where d: C (M) Ω 1 (M)isthestandarddifferential, then d d: C (M) C (M)iscalledthescalarLaplacianandwillbedenoted by M. In what follows, we use the notations (1.2) (Re V)(x):= V(x)+(V(x)) 2, (Im V)(x):= V(x) (V(x)) 2i, x M, where i= 1and(V(x)) denotestheadjointofthelinearoperator V(x): E x E x (inthesenseoflinearalgebra). By(1.2),forall x M,(Re V)(x)and(ImV)(x)areself-adjointlinearoperators E x E x,andwehavethefollowingdecomposition: V(x)=(Re V)(x)+i(Im V)(x). Forevery x M,wehavethefollowingdecomposition: (1.3) (Re V)(x)=(Re V) + (x) (Re V) (x). Here(Re V) + (x)=p + (x)(re V)(x),where P + (x):= χ [0,+ ) ((Re V)(x)),and (Re V) (x)= P (x)(re V)(x),where P (x):= χ (,0) ((Re V)(x)). Here χ A denotes the characteristic function of the set A. Wemakethefollowingassumptionon V.

On m-sectorial Schrödinger-type operators... 93 Assumption A. (i) (Re V) + L 1 loc (EndE),(Re V) L 1 loc (EndE)and(ImV) L1 loc (EndE). (ii) Thereexistsaconstant L >0suchthatforall u L 2 (E)andall x M, (1.4) (Im V)(x) u(x) 2 L (Re V) + (x)u(x), u(x), where (Im V)(x) denotesthenormoftheoperator(im V)(x): E x E x, u(x) denotesthenorminthefiber E x and, denotestheinnerproduct in E x. 1.2Sobolevspace W 1,2 (E). By W 1,2 (E)wewilldenotethecompletionofthe space C c (E)withrespecttothenorm 1definedbythescalarproduct (u, v) 1 := (u, v) + ( u, v) u, v C c (E). By W 1,2 (E)wewilldenotethedualof W 1,2 (E). 1.3 Quadratic forms. In what follows, all quadratic forms are considered in thehilbertspace L 2 (E). 1.By h 0 wedenotethequadraticform (1.5) h 0 (u)= u 2 dµ withthedomaind(h 0 )=W 1,2 (E) L 2 (E). Thequadraticform h 0 isnonnegative,denselydefined(since C c (E) D(h 0 ))andclosed(seesection1.2). 2.By h 1 wedenotethequadraticform (1.6) h 1 (u)= (Re V) + u, u +i (Im V)u, u dµ with the domain (1.7) D(h 1 )= { u L 2 (E): (Re V) + u, u +i (Im V)u, u dµ <+ }. Here, denotesthefiberwiseinnerproductin E x. Inwhatfollows,wewilldenoteby h 1 (, )thecorrespondingsesquilinearform obtainedviapolarizationidentityfrom h 1. Thequadraticform h 1 issectorial.indeed,bytheinequalities (1.8) (Im V)u(x), u(x) (Im V)(x)u(x) u(x) (Im V)(x) u(x) 2 andby(1.4),forall u D(h 1 ),thevaluesof h 1 (u)lieinasectorof Cwithvertex γ=0. Theform h 1 isdenselydefinedsince,by(i)ofassumptiona,wehave

94 O. Milatovic Cc (E) D(h 1). Theform h 1 isclosed. Indeed,byTheoremVI.1.11in[6],it sufficestoshowthatthepre-hilbertspaced(h 1 )withtheinnerproduct (u, v) h1 =(Re h 1 )(u, v)+(u, v)= (Re V) + u, v dµ+(u, v), iscomplete.here(, )denotestheinnerproductin L 2 (E)and(Re h 1 )(, )denotes therealpartofthesesquilinearform h 1 (, )(seethedefinitionbelowtheequation (1.9) in Section VI.1.1 of[6]). By(1.7),(1.8)and(1.4),itfollowsthatD(h 1 )isthesetofall u L 2 (E)such that u 2 h 1 <+,where h1 denotesthenormcorrespondingtotheinner product(, ) h1.byexamplevi.1.15in[6],itfollowsthatd(h 1 )iscomplete. 3.By h 2 wedenotethequadraticform (1.9) h 2 (u)= (Re V) u, u dµ with the domain (1.10) D(h 2 )= { u L 2 (E): (Re V) u, u dµ <+ }, where, denotesthefiberwiseinnerproductin E x. Theform h 2 isdenselydefinedbecause, by(i)ofassumptiona,wehave C c (E) D(h 2 ).Moreover,sinceforall x M,theoperator(Re V) (x): E x E x isself-adjoint,itfollowsthatthequadraticform h 2 issymmetric. Wemakethefollowingassumptionon h 2. AssumptionB. Assumethat h 2 is h 0 -boundedwithrelativebound b <1,i.e. (i) D(h 2 ) D(h 0 ), (ii) thereexistconstants a 0and0 b <1suchthat (1.11) h 2 (u) a u 2 + b h 0 (u), forall u D(h 0 ), where denotesthenormin L 2 (E). Remark1.4.If(M, g)isamanifoldofboundedgeometry,assumptionbholdsif (Re V) L p (EndE),where p=n/2for n 3, p >1for n=2,and p=1for n=1.fortheproof,see,forexample,theproofofremark2.1in[7]. 1.5Arealizationof H V in L 2 (E). Wedefineanoperator Sin L 2 (E)bythe formula Su=H V uonthedomain (1.12) { (Re u W 1,2 (E): V) + u, u +i (Im V)u, u dµ <+ and H V u L 2 (E) }.

On m-sectorial Schrödinger-type operators... 95 Wewilldenotethesetin(1.12)byDom(S). Remark1.6.Forall u D(h 0 )=W 1,2 (E)wehave u W 1,2 (E). From Corollary2.11belowitfollowsthatforall u W 1,2 (E) D(h 1 ),wehave V u L 1 loc (E). Thus H V uin(1.12)isadistributionalsectionof E,andthecondition H V u L 2 (E)makessense. Remark1.7.By(1.4)andby(1.8),thesetDom(S)in(1.12)isequalto { u W 1,2 (E): (Re V) + u, u dµ <+ and H V u L 2 (E) }. Wenowstatethemainresult. Theorem 1.8. Assumethat(M, g)isamanifoldofboundedgeometrywith positive C -boundedmeasure dµ, EisaHermitianvectorbundleofbounded geometryover M,and isac -boundedhermitianconnectionon E.Suppose thatassumptionsaandbhold.thentheoperator Sis m-sectorial. Remark1.9.Theorem1.8extendsaresultofT.Kato;seeTheoremVI.4.6(a)in[6] (seealsoremark5(a)in[5])whichwasprovenfortheoperator +V,where isthestandardlaplacianon R n withthestandardmetricandmeasure,and V L 1 loc (Rn )isasinassumptionsaandbabove(withimv =0).Theorem1.8 also extends the result in[7] which establishes the self-adjointness of a realization in L 2 (E)of H V = +V onmanifold(m, g)with dµ, E,and asinthe hypothesesoftheorem1.8,and V = V 1 + V 2,where0 V 1 L 1 loc (EndE)and 0 V 2 L 1 loc (EndE)arelinearself-adjointbundleendomorphismssatisfying AssumptionsAandB(withImV =0). 2. ProofofTheorem1.8 WeadopttheargumentsfromSectionVI.4.3in[6]tooursettingwiththehelp of a more general version of Kato s inequality(2.1). 2.1 Kato s inequality. We begin with the following variant of Kato s inequality for Bochner Laplacian(for the proof, see Theorem 5.7 in[2]). Lemma2.2. Assumethat(M, g)isariemannianmanifold.assumethat Eisa Hermitianvectorbundleover Mand isahermitianconnectionon E.Assume that w L 1 loc (E)and w L 1 loc (E).Then (2.1) M w Re w,sign w, where sign w(x) = { w(x) if w(x) 0, w(x) 0 otherwise. Remark 2.3. The original version of Kato s inequality was proven in Kato[3].

96 O. Milatovic 2.4 Positivity. In what follows, we will use the following lemma whose proof is giveninappendixbof[2]. Lemma 2.5. Assumethat(M, g)isamanifoldofboundedgeometrywitha smooth positive measure dµ. Assume that ( b+ M ) u = ν 0, u L 2 (M), where b >0, M = d disthescalarlaplacianon M,andtheinequality ν 0 meansthat ν isapositivedistributionon M,i.e.(ν, φ) 0forany0 φ C c (M). Then u 0(almost everywhere or, equivalently, as a distribution). Remark2.6.ItisnotknownwhetherLemma2.5holdsif Misanarbitrarycomplete Riemannian manifold. For more details about difficulties in the case of arbitrary complete Riemannian manifolds, see Appendix B of[2]. Lemma2.7. Thequadraticform h:=(h 0 +h 1 )+h 2 isdenselydefined,sectorial and closed. Proof:Since h 0 and h 1 aresectorialandclosed,itfollowsbytheoremvi.1.31 from[6]that h 0 +h 1 issectorialandclosed.by(i)ofassumptionbitfollowsthat D(h 2 ) D(h 0 ) D(h 1 ),andby(1.11),(1.5),and(1.6),thefollowinginequality holds: h 2 (u) a u 2 + b h 0 (u)+h 1 (u), forall u D(h 0 ) D(h 1 ), where denotesthenormin L 2 (E),and a 0and0 b <1areasin(1.11). Thusthequadraticform h 2 is(h 0 +h 1 )-boundedwithrelativebound b <1.Since h 0 + h 1 isaclosedsectorialform,bytheoremvi.1.33from[6],itfollowsthat h=(h 0 + h 1 )+h 2 isaclosedsectorialform. Since Cc (E) D(h 0 ) D(h 1 ) D(h 2 ),itfollowsthat hisdenselydefined. In what follows, h(, ) will denote the corresponding sesquilinear form obtained from h via polarization identity. 2.8 m-sectorial operator H associated with h. Since h is a densely defined, closedandsectorialformin L 2 (E),byTheoremVI.2.1from[6]thereexistsan m-sectorialoperator Hin L 2 (E)suchthat (i) Dom(H) D(h)and h(u, v)=(hu, v), (ii) Dom(H)isacoreof h, (iii) if u D(h), w L 2 (E),and forall u Dom(H)and v D(h), h(u, v)=(w, v) holdsforevery vbelongingtoacoreof h,then u Dom(H)and Hu=w. The operator H is uniquely determined by condition(i). We will also use the following lemma.

On m-sectorial Schrödinger-type operators... 97 Lemma2.9. Assumethat0 T L 1 loc (EndE)isalinearself-adjointbundle map.assumealsothat u Q(T),where Q(T)={u L 2 (E): Tu, u L 1 (M)}. Then Tu L 1 loc (E). Proof:Byaddingaconstantwecanassumethat T 1(inoperatorsense). Assumethat u Q(T). Wechoose(inameasurableway)anorthogonal basisin eachfiber E x anddiagonalize1 T(x) End(E x ) toget T(x) = diag(c 1 (x), c 2 (x),..., c m (x)), where 0 < c j L 1 loc (M), j = 1,2,..., m and m=dime x. Let u j (x)(j=1,2,..., m)bethecomponentsof u(x) E x withrespectto thechosenorthogonalbasisof E x.thenforall x M Tu, u = m c j (x) u j (x) 2. j=1 Since u Q(T),weknowthat0< Tu, u dµ <+. Since c j >0,itfollows that c j u j 2 L 1 (M),forall j=1,2,...,m. Now,forall x Mand j=1,2,, m (2.2) 2 c j u j =2 c j u j c j + c j u j 2. Therighthandsideof(2.2)isclearlyin L 1 loc (M).Therefore c ju j L 1 loc (M). But(Tu)(x)hascomponents c j (x)u j (x)(j=1,2,...,m)withrespecttochosenbasesof E x.therefore Tu L 1 loc (E),andthelemmaisproven. Corollary2.10. If u D(h 1 ),then((re V) + + i(im V))u L 1 loc (E). Proof:Let u D(h 1 ).Then (Re V) + u, u L 1 (M),and,hence,byLemma2.9 weget(re V) + u L 1 loc (E).By(1.4)weobtain (Im V) u 2 L 1 (M).Sincefor all x Mwehave 2 (Im V)(x)u(x) 2 (Im V)(x) u(x) (Im V)(x) + (Im V)(x) u(x) 2, and since, by AssumptionA, (Im V) L 1 loc (M), it followsthat (ImV)u L 1 loc (E),andthecorollaryisproven. Corollary2.11. If u D(h),then V u L 1 loc (E). Proof: Let u D(h) = D(h 0 ) D(h 1 ). ByCorollary2.10it followsthat ((Re V) + + i(im V))u L 1 loc (E).SinceD(h) D(h 2)andsince(Re V) (x) 0 asanoperator E x E x,bylemma2.9wehave(re V) u L 1 loc (E). Thus V u L 1 loc (E),andthecorollaryisproven.

98 O. Milatovic Lemma 2.12. The following operator relation holds: H S. Proof:Wewillshowthatforall u Dom(H),wehave Hu=H V u. Let u Dom(H).Byproperty(i)ofSection2.8wehave u D(h);hence,by Corollary2.11weget V u L 1 loc (E).Then,forany v C c (E),wehave (2.3) (Hu, v)=h(u, v)=( u, v)+ V u, v dµ, where(, )denotesthe L 2 -innerproduct. Thefirstequalityin(2.3)holdsbyproperty(i)fromSection2.8,andthesecond equality holds by definition of h. Hence,usingintegrationbypartsinthefirsttermontherighthandsideof thesecondequalityin(2.3)(see,forexamplelemma8.8from[2]),weget (2.4) (u, v)= Hu V u, v dµ, forall v C c (E). Since V u L 1 loc (E)and Hu L2 (E),itfollowsthat(Hu V u) L 1 loc (E),and (2.4)implies u=hu V u(asdistributionalsectionsof E).Therefore, u+vu=hu, andthisshowsthat Hu=H V uforall u Dom(H). Nowbydefinitionof SitfollowsthatDom(H) Dom(S)and Hu=Sufor all u Dom(H).Therefore H S,andthelemmaisproven. Lemma2.13. C c (E)isacoreofthequadraticform h 0+ h 1. Proof:Itsufficestoshow(seeTheoremVI.1.21in[6]andtheparagraphabove theequation(1.31)insectionvi.1.3of[6])that C c (E)isdenseintheHilbert spaced(h 0 + h 1 )=D(h 0 ) D(h 1 )withtheinnerproduct (u, v) h0 +h 1 := h 0 (u, v)+(re h 1 )(u, v)+(u, v), where h 0 (, )denotesthesesquilinearformcorrespondingto h 0 viapolarization identityand(re h 1 )denotestherealpartofthesesquilinearform h 1 (, ). Let u D(h 0 + h 1 )and(u, v) h0 +h 1 =0forall v C c (E).Wewillshowthat u=0. We have (2.5) 0 = h 0 (u, v)+(re h 1 )(u, v)+(u, v) = (u, v)+ (Re V) + u, v dµ+(u, v).

On m-sectorial Schrödinger-type operators... 99 Hereweusedintegrationbypartsinthefirsttermontherighthandsideofthe second equality. Since u D(h 0 +h 1 ) D(h 1 ),itfollowsthat (Re V) + u, u +i (ImV)u, u L 1 (M). Hence (Re V) + u, u L 1 (M). ByLemma2.9weget(Re V) + u L 1 loc (E).From(2.5)wegetthefollowingdistributionalequality: (2.6) u=( (Re V) + 1)u. From(2.6)wehave u L 1 loc (E).ByLemma2.2andby(2.6),weobtain (2.7) M u Re u,signu = ((Re V) + +1)u,signu u. Thelastinequalityin(2.7)holdssince(Re V) + (x) 0asanoperator E x E x. Therefore, (2.8) ( M +1) u 0. ByLemma2.5,itfollowsthat u 0.So u=0,andthelemmaisproven. Lemma2.14. C c (E)isacoreofthequadraticform h=(h 0+ h 1 )+h 2. Proof: Sincethequadraticform h 2 is(h 0 + h 1 )-bounded,thelemmafollows immediately from Lemma 2.13. 3. ProofofTheorem1.8 ByLemma2.12wehave H S,soitisenoughtoshowthatDom(S) Dom(H). Let u Dom(S).BydefinitionofDom(S)inSection1.5,wehave u D(h 0 ) D(h 2 )and u D(h 1 ).Hence u D(h). Forall v C c (E)wehave h(u, v)=h 0 (u, v)+h 1 (u, v)+h 2 (u, v)=(u, v)+ V u, v dµ=(h V u, v). Thelastequalityholdssince H V u=su L 2 (E). ByLemma2.14itfollows that C c (E)isaformcoreof h. Nowfromproperty(iii)ofSection2.8wehave u Dom(H)with Hu=H V u.thisconcludestheproofofthetheorem. References [1] Aubin T., Some Nonlinear Problems in Riemannian Geometry, Springer-Verlag, Berlin, 1998. [2] Braverman M., Milatovic O., Shubin M., Essential self-adjointness of Schrödinger type operators on manifolds, Russian Math. Surveys 57(2002), no. 4, 641 692. [3] Kato T., Schrödinger operators with singular potentials, Israel J. Math. 13(1972), 135 148.

100 O. Milatovic [4] Kato T., A second look at the essential selfadjointness of the Schrödinger operators, Physical reality and mathematical description, Reidel, Dordrecht, 1974, pp. 193 201. [5] Kato T., On some Schrödinger operators with a singular complex potential, Ann. Scuola Norm.Sup.PisaCl.Sci(4)5(1978),105 114. [6] Kato T., Perturbation Theory for Linear Operators, Springer-Verlag, New York, 1980. [7] Milatovic O., Self-adjointness of Schrödinger-type operators with singular potentials on manifolds of bounded geometry, Electron. J. Differential Equations, No. 64(2003), 8pp (electronic). [8] Reed M., Simon B., Methods of Modern Mathematical Physics I, II: Functional Analysis. Fourier Analysis, Self-adjointness, Academic Press, New York e.a., 1975. [9] Shubin M.A., Spectral theory of elliptic operators on noncompact manifolds, Astérisque no. 207(1992), 35 108. [10] Taylor M., Partial Differential Equations II: Qualitative Studies of Linear Equations, Springer-Verlag, New York e.a., 1996. Department of Mathematics, University of Toledo, Toledo, OH 43606-3390, USA E-mail: Ognjen.Milatovic@utoledo.edu (Received September 9, 2003)