• No results found

Minimally thin sets associated with the stationary Schrödinger operator

N/A
N/A
Protected

Academic year: 2020

Share "Minimally thin sets associated with the stationary Schrödinger operator"

Copied!
9
0
0

Loading.... (view fulltext now)

Full text

(1)

R E S E A R C H

Open Access

Minimally thin sets associated with the

stationary Schrödinger operator

Tao Zhao

*

*Correspondence: [email protected]

College of Mathematics and Information Science, Henan University of Economics and Law, Zhengzhou, 450000, P.R. China

Abstract

This paper gives some new criteria fora-minimally thin sets at infinity with respect to the Schrödinger operator in a cone, which supplement the results obtained by Long-Gao-Deng.

MSC: 31B05; 31B10

Keywords: minimally thin set; Schrödinger operator; Greena-potential

1 Introduction and results

LetRandR+ be the set of all real numbers and the set of all positive real numbers, re-spectively. We denote byRn(n≥) then-dimensional Euclidean space. A point inRnis denoted byP= (X,xn),X= (x,x, . . . ,xn–). The Euclidean distance between two pointsP andQinRnis denoted by|PQ|. Also|PO|withOthe origin ofRnis simply denoted

by|P|. The boundary and the closure of a setSinRnare denoted bySandS, respectively.

We introduce a system of spherical coordinates (r,),= (θ,θ, . . . ,θn–), inRnwhich are related to Cartesian coordinates (x,x, . . . ,xn–,xn) byxn=rcosθ.

LetDbe an arbitrary domain inRnandA

adenote the class of nonnegative radial

po-tentialsa(P),i.e.≤a(P) =a(r),P= (r,)∈D, such thataLbloc(D) with someb>n/ if

n≥ and withb=  ifn=  orn= .

Ifa∈Aa, then the stationary Schrödinger operator

Scha= –+a(P)I= ,

whereis the Laplace operator andIis the identical operator, can be extended in the usual way from the spaceC∞(D) to an essentially self-adjoint operator onL(D) (see [, Ch. ]). We will denote itScha as well. This last one has a Greena-functionGaD(P,Q).

HereGaD(P,Q) is positive onDand its inner normal derivative∂GaD(P,Q)/∂nQ≥, where

/∂nQdenotes differentiation atQalong the inward normal intoD.

We call a functionu≡–∞that is upper semi-continuous inDa subfunction with re-spect to the Schrödinger operatorSchaif it values belong to the interval [–∞,∞) and at

each pointPDwith  <r<r(P) the generalized mean-value inequality (see [])

u(P)≤

S(P,r)

u(Q)∂G

a

B(P,r)(P,Q)

∂nQ

(Q)

(2)

is satisfied, whereGBa(P,r)(P,Q) is the Greena-function ofScha inB(P,r) and(Q) is a

surface measure on the sphereS(P,r) =∂B(P,r). If –uis a subfunction, then we callua superfunctions (with respect to the Schrödinger operatorScha). If a functionuis both

sub-function and supersub-function, it is, clearly, continuous and is called ana-harmonic function (with respect to the Schrödinger operatorScha).

The unit sphere and the upper half unit sphere in Rnare denoted bySn–andSn– + , respectively. For simplicity, a point (,) on Sn– and the set {; (,)} for a set

,Sn–, are often identified withand, respectively. For two setsR + and

Sn–, the set {(r,)Rn;r, (,)}inRn is simply denoted by×. By

Cn(), we denote the setRinRnwith the domainonSn–. We call it a cone. We denote the setI×with an interval onRbyCn(;I).

From now on, we always assume D=Cn(). For the sake of brevity, we shall write

Ga(P,Q) instead ofGCan()(P,Q). Throughout this paper, letcdenote various positive

con-stants, because we do not need to specify them.

Letbe a domain onSn–with smooth boundary. Consider the Dirichlet problem ( n+λ)ϕ=  on,

ϕ=  on,

where nis the spherical part of the Laplace operatorn

n=

n– 

r ∂r+

∂r +

n

r .

We denote the least positive eigenvalue of this boundary value problem byλand the nor-malized positive eigenfunction corresponding toλbyϕ(). In order to ensure the exis-tence ofλand a smoothϕ(), we put a rather strong assumption on: ifn≥, then

is aC,α-domain ( <α< ) onSn–surrounded by a finite number of mutually disjoint closed hypersurfaces (e.g.see [, pp.-] for the definition ofC,α-domain).

For any (,)∈, we have (see [, pp.-])

c–()≤δ(P)≤crϕ(), ()

whereP= (r,)∈Cn() andδ(P) =dist(P,∂Cn()).

We study solutions of an ordinary differential equation,

Q (r) –n– 

r Q(r) +

λ r+a(r)

Q(r) = ,  <r<∞. ()

It is well known (see, for example, []) that if the potentiala∈Aa, then equation () has

a fundamental system of positive solutions {V,W} such thatV is nondecreasing with (see [])

≤V(+)≤V(r) ∞ asr→+∞,

andWis monotonically decreasing with

(3)

We will also consider the classBa, consisting of the potentialsa∈Aasuch that the finite

limitlimr→∞ra(r) =k∈[,∞) exists, and moreover,r–|ra(r) –k| ∈L(,∞). Ifa∈Ba,

then the (sub)superfunctions are continuous (see []). In the rest of this paper, we assume thata∈Baand we shall suppress this assumption for simplicity.

Denote

ι±k = –n±

(n– )+ (k+λ)

 ,

then the solutions to equation () have the asymptotic (see [])

c–+kV(r)crι+k, c–rιkW(r)crιk asr→ ∞. ()

It is well known that the Martin boundary ofCn() is the set∂Cn()∪ {∞}, each point

of which is a minimal Martin boundary point. ForPCn() andQ∂Cn()∪ {∞}, the

Martin kernel can be defined byMa

(P,Q). If the reference pointPis chosen suitably, then we have

Ma(P,∞) =V(r)ϕ() and Ma(P,O) =cW(r)ϕ() ()

for anyP= (r,)∈Cn().

In [], Long-Gao-Deng introduce the notations ofa-thin (with respect to the Schrö-dinger operatorScha) at a point, a-polar set (with respect to the Schrödinger operator

Scha) anda-minimal thin sets at infinity (with respect to the Schrödinger operatorScha),

which generalized earlier notations obtained by Brelot and Miyamoto (see [, ]). A setH

inRnis said to bea-thin at a pointQif there is a fine neighborhoodEofQwhich does not intersectH\{Q}. OtherwiseHis said to be nota-thin atQonCn(). A setHinRnis called

a polar set if there is a superfunctionuon some open setEsuch thatH⊂ {PE;u(P) =∞}. A subsetHofCn() is said to bea-minimal thin atQ∂Cn()∪ {∞}onCn(), if there

exists a pointPCn() such that

ˆ

RHMa

(·,Q)(P)=M

a

(P,Q),

whereRˆH

Ma(·,Q)is the regularized reduced function ofMa(·,Q) relative toH(with respect to the Schrödinger operatorScha).

LetHbe a bounded subset ofCn(). ThenRˆHMa

(·,∞)(P) is bounded onCn() and hence

the greatest a-harmonic minorant ofRˆH

Ma(·,∞) is zero. When byG a

μ(P) we denote the Greena-potential with a positive measureμonCn(), we see from the Riesz

decomposi-tion theorem (see [, Theorem ]) that there exists a unique positive measureλaHonCn()

such that (see [, p.])

ˆ

RHMa

(·,∞)(P) =G

a

λ

a H(P)

for anyPCn() andλaHis concentrated onIH, where

IH=

PCn();His nota-thin atP

(4)

The Green a-energyγa(H) (with respect to the Schrödinger operatorScha) ofλaH is

defined by

γa(H) =

Cn()

GaλaHdλaH.

Also, we can define a measureσa onCn()

σa(H) =

H

Ma(P,∞)

δ(P)

dP.

Recently, Long-Gao-Deng (see [, Theorem .]) gave a criterion that characterizes

a-minimally thin sets at infinity in a cone.

Theorem A A subset H of Cn()is a-minimally thin at infinity on Cn()if and only if

j=

γa(Hj)Wj

V– j<∞,

where Hj=HCn(; [j, j+))and j= , , , . . . .

In this paper, we shall obtain a series of new criteria fora-minimally thin sets at infinity onCn(), which complemented Theorem A by the way completely different from theirs.

Our results are essentially based on Ren and Su (see [, ]). First we have the following.

Theorem  The following statements are equivalent.

(I) A subsetHofCn()isa-minimally thin at infinity onCn().

(II) There exists a positive superfunctionv(P)onCn()such that

inf

PCn()

v(P)

Ma

(P,∞)

=  ()

and

HPCn();v(P)≥Ma(P,∞)

.

(III) There exists a positive superfunctionv(P)onCn()such that even if

v(P)≥cMa(P,∞)for anyPH,there existsP∈Cn()satisfying

v(P) <cMa(P,∞).

Next we shall state Theorem , which is the main result in this paper.

Theorem  If a subset H of Cn()is a-minimally thin at infinity on Cn(),then we have

H

dP

(5)

2 Lemmas

In our discussions, the following estimate for the Greena-potentialGa(P,Q) is funda-mental, as follows from [].

Lemma 

c–V(r)W(t)ϕ()ϕ()≤Ga(P,Q)≤cV(r)W(t)ϕ()ϕ()

for any P= (r,)∈Cn()and any Q= (t,)∈Cn()satisfying t≥r.

Lemma  If H is a bounded Borel subset of Cn(),then

σa(H)≤a(H).

Proof For anyPRn\C

n() and any positive numberr> , there exists a positive constant

csuch that

Cap P+r–(QP)∈Rn;QB(P,r)∩ Rn\Cn()

c

from [, p.], whereCapdenotes the Newtonian capacity. Then there exists a positive constantcdepending only oncandnsuch that

Cn()

(P)

δ(P)

dPc

Cn()

(P)dP ()

for every(P)∈C∞ (Cn()) (see [, Theorem ]).

It is well known that the Greena-energy also can be represented as (see [, p.])

γa(H) =

Cn()

GaλaH(P)

dP. ()

From equation () and Lemma  we have

Cn()

GaλaH(P)

δ(P)

dP<∞. ()

From equations () and () we obtainGaλaH(P)∈, where

=

fLloc Cn()

;∇fLCn()

,δ–fLCn()

equipped with the norm

f= ∇f

L(Cn())+δ–f

L(Cn())

 ,

and furtherGa

λaH(P)∈, where denotes the closure ofC∞(Cn()) in. Thus we obtain from equation () (see [, p.])

Cn()

GaλaH(P)

δ(P)

dPc

Cn()

(6)

SinceGaλaH=Ma(·,∞) quasi everywhere onHand hence a.e. onH, we have from equa-tion ()

γa(H)≥c–

Cn()

Ga

λaH(P)

δ(P)

dP

c–

Cn()

Ma

(P,∞)

δ(P)

dP

=c–σa(H),

which gives the conclusion of Lemma .

3 Proof of Theorem 1

We shall show that (II) follows from (I). Since

ˆ

RHMja

(·,∞)(Q) =M

a

(Q,∞)

for anyQIHjandλHj is concentrated onIHj, we have

γa(Hj) =

IHj

Ma(Q,∞)dλaHj(Q)

Vj

IHj

ϕ()dλaH j(Q)

for anyQ= (t,)∈Cn() and hence from Lemma 

ˆ

RHMja

(·,∞)(P)≤cV(r)ϕ()

IHj

W(t)ϕ()dλaH j(Q)

cV(r)ϕ()WjV– a(Hj) ()

for anyP= (r,)∈Cn() and any integerjsatisfying j≥r.

If we define a measureμonCn() by

μ= ∞

j=

λaH j,

then

Gaμ(P) = ∞

j=

ˆ

RHMja (·,∞)(P).

From equation (), (I), and Theorem A, we know thatGaμ(P) is a finite superfunction onCn() and

Gaμ(P)≥ ˆR

Hj

(7)

for anyP= (r,)∈IHj (j= , , , , . . .) and from Lemma 

Gaμ(P)≥cV(r)ϕ()

for anyP= (r,)∈Cn(; (, )) and

c=c–

Cn(;[r,∞))

W(t)ϕ()(Q).

If we setH =∞j=–IHj, where

H–=HCn ; (, )

, ()

andc=min{c, }, then

HP= (r,)∈Cn();Gaμ(P)≥cV(r)ϕ()

andH is equal toHexcept a polar setH. If we define a positive measureηonCn() such

thatGaμis identically +∞onHand define a measureνonCn() byν=c– (μ+η), then

HP= (r,)∈Cn();Gaν(P)≥V(r)ϕ()

.

If we putv(P) =Ga

ν(P), then this shows thatv(P) is the function required in (II). Now we shall show that (III) follows from (II). Letv(P) be the function in (II). It follows thatv(P)≥Ma

(P,∞) for anyPH. On the other hand, from equation () we can find a pointP∈Cn() such thatv(P) <Ma(P,∞). Thereforev(P) satisfies (III) withc= .

Finally, we shall prove that (I) follows from (III). Letv(P) be the function in (III). If we put

inf

PCn() v(P)

Ma(P,∞)

=c(∞,v)

and

u(P) =v(P) –c(∞,v)Ma(P,∞),

then we have

inf

PCn() u(P)

Ma(P,∞) = ,

wherec(∞,v) is a positive constant depending only on∞andv. Since there existsP∈

Cn() satisfyingv(P) <cMa(P,∞), we note thatc>c(∞,v). Now we obtainu(P)≥ (c–c(∞,v))Ma(P,∞) for anyPH. Hence by a result of [, p.],H isa-minimally thin at infinity onCn() with respect to the Schrödinger operator, which is the statement

(8)

4 Proof of Theorem 2 First of all, we remark that

H

dP

( +|P|)n =

H–

dP

( +|P|)n+

j=

Hj dP

( +|P|)n

≤ |H–|+ ∞

j=

–jn|Hj|, ()

whereH–is the set in equation () and|Hj|is then-dimensional Lebesgue measure ofHj.

We have from equations () and ()

σa(Hj) =

Hj

Ma

(P,∞)

δ(P)

dP

c

Hj

V(r)ϕ()

()

dP

c

Hj

rι+k–dP

c

Hj

j(ι+k–)dP

=cj(ι+k–)|Hj|.

By using Lemma , we obtain

γa(Hj)≥c–σa(Hj)≥cj(ι

+

k–)|Hj|. ()

IfHisa-minimally thin at infinity onCn(), then from Theorem A, equations (), (),

and (), we have

H

dP

( +|P|)n ≤ |H–|+c

j=

j(ι+k–)|Hj|W jV– j

≤ |H–|+c

j=

γa(Hj)Wj

V– j <∞,

which is the conclusion of Theorem .

Competing interests

The author declares that there is no conflict of interests regarding the publication of this article.

Acknowledgements

This work was supported by the National Natural Science Foundation of China under Grants Nos. 11301140 and U1304102. The author would like to thank two anonymous referees for numerous insightful comments and suggestions, which have greatly improved the paper.

(9)

References

1. Levin, B, Kheyfits, A: Asymptotic behavior of subfunctions of time-independent Schrödinger operator. In: Some Topics on Value Distribution and Differentiability in Complex andP-Adic Analysis, chap. 11, pp. 323-397. Science Press, Beijing (2008)

2. Gilbarg, D, Trudinger, NS: Elliptic Partial Differential Equations of Second Order. Springer, Berlin (1977) 3. Courant, R, Hilbert, D: Methods of Mathematical Physics, vol. 1. Interscience, New York (2008)

4. Verzhbinskii, GM, Maz’ya, VG: Asymptotic behavior of solutions of elliptic equations of the second order close to a boundary. I. Sib. Math. J.12, 874-899 (1971)

5. Long, PH, Gao, ZQ, Deng, GT: Criteria of Wiener type for minimally thin sets and rarefied sets associated with the stationary Schrödinger operator in a cone. Abstr. Appl. Anal.2012, Article ID 453891 (2012)

6. Simon, B: Schrödinger semigroups. Bull. Am. Math. Soc.7, 447-526 (1982)

7. Brelot, M: On Topologies and Boundaries in Potential Theory. Lecture Notes in Mathematics, vol. 175. Springer, Berlin (1971)

8. Miyamoto, I, Yoshida, H: Two criteria of Wiener type for minimally thin sets and rarefied sets in a cone. J. Math. Soc. Jpn.54, 487-512 (2002)

9. Ren, YD: Solving integral representations problems for the stationary Schrödinger equation. Abstr. Appl. Anal.2013, Article ID 715252 (2013)

10. Su, BY: Dirichlet problem for the Schrödinger operator in a half space. Abstr. Appl. Anal.2012, Article ID 578197 (2012) 11. Lewis, IL: Uniformly fat sets. Trans. Am. Math. Soc.308, 177-196 (1988)

12. Long, PH: The Characterizations of Exceptional Sets and Growth Properties in Classical or Nonlinear Potential Theory. Dissertation of Beijing Normal University, Beijing Normal University, Beijing, China (2012)

13. Ancona, A: On strong barriers and an inequality of Hardy for domains inRn. J. Lond. Math. Soc.34, 274-290 (1986)

10.1186/1029-242X-2014-67

References

Related documents