HAL Id: hal-01819321
https://hal.archives-ouvertes.fr/hal-01819321
Submitted on 20 Jun 2018
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
SCHRODINGER OPERATORS WITH NEGATIVE POTENTIALS AND LANE-EMDEN DENSITIES
Lorenzo Brasco, Giovanni Franzina, Berardo Ruffini
To cite this version:
Lorenzo Brasco, Giovanni Franzina, Berardo Ruffini. SCHRODINGER OPERATORS WITH NEG-
ATIVE POTENTIALS AND LANE-EMDEN DENSITIES. Journal of Functional Analysis, Elsevier,
2018. �hal-01819321�
WITH NEGATIVE POTENTIALS AND LANE-EMDEN DENSITIES
LORENZO BRASCO, GIOVANNI FRANZINA, AND BERARDO RUFFINI
Abstract. We consider the Schr¨odinger operator−∆ +V for negative potentialsV, on open sets with positive first eigenvalue of the Dirichlet-Laplacian. We show that the spectrum of−∆ +V is positive, provided thatV is greater than a negative multiple of the logarithmic gradient of the solution to the Lane-Emden equation−∆u=uq−1 (for some 1≤q <2). In this case, the ground state energy of −∆ +V is greater than the first eigenvalue of the Dirichlet-Laplacian, up to an explicit multiplicative factor. This is achieved by means of suitable Hardy-type inequalities, that we prove in this paper.
Contents
1. Introduction 1
1.1. Foreword 1
1.2. Aim of the paper 3
1.3. Main results 4
1.4. Plan of the paper 5
2. Preliminaries 5
2.1. Notation 5
2.2. Lane-Emden densities: bounded sets 6 2.3. Lane-Emden densities: general sets 8
3. Hardy-Lane-Emden inequalities 9
4. Sobolev embeddings and densities 11 5. Hardy-Lane-Emden inequalities for sets with positive
spectrum 16
6. Lower bounds for the ground state energy 17
7. Applications 20
7.1. N−dimensional ball 20
7.2. An infinite slab 21
7.3. A rectilinear wave-guide 22
Appendix A. A local L∞ estimate for Lane-Emden
densities 25
References 29
1. Introduction
1.1. Foreword. Let V ∈ L
2loc( R
N) be a real-valued potential such that V ≤ 0 and let us consider the Schr¨ odinger operator H
V:= − ∆ + V , acting on the domain
D( H
V) := H
2(R
N) ∩ { u ∈ L
2(R
N) : V u ∈ L
2(R
N) } .
2010Mathematics Subject Classification. 35P15, 47A75, 49S05.
Key words and phrases. Schr¨odinger operators, ground state energy, Hardy inequalities, Lane-Emden equation.
1
Observe that the hypothesis V ∈ L
2loc(R
N) entails the inclusion C
0∞( R
N) ⊂ D( H
V),
thus D( H
V) is dense in L
2( R
N). The operator H
V: D( H
V) → L
2( R
N) is symmetric and self- adjoint as well, thanks to the fact that V is real-valued (see [16, Example p. 68]). The spectrum of H
Vis the set
σ( H
V) = R \ ρ( H
V),
where ρ( H
V) is the resolvent set of H
V, defined as the collection of real numbers λ such that H
V− λ is bijective and its inverse is a bounded linear operator.
A distinguished subset of σ( H
V) is given by the collection of those λ such that the kernel of H
V− λ is nontrivial. In this case, the stationary Schr¨ odinger equation
(1.1) H
Vu = λ u,
admits a nontrivial solution u ∈ D( H
V). Whenever this happens, λ is called an eigenvalue of the Schr¨ odinger operator. Correspondingly, the solution is said to be an eigenfunction corresponding to λ.
The operator H
Vcomes with the associated quadratic form ϕ 7→ Q
V(ϕ) =
ˆ
RN
|∇ ϕ |
2dx + ˆ
RN
V ϕ
2dx, ϕ ∈ D( H
V).
From classical Spectral Theory, we have (see [16, Theorem 2.20])
(1.2) inf σ( H
V) = inf
ϕ∈D(HV)
Q
V(ϕ) : ˆ
RN
ϕ
2dx = 1
. We call such a value ground state energy of H
V.
This quantity is important in classical Quantum Mechanics, since it is the lowest energy that a particle in R
Ninteracting with the force field generated by the potential V can attain (and which will eventually attain by emitting energy). From a mathematical point of view, we observe that the stationary Schr¨ odinger equation (1.1) is precisely the Euler-Lagrange equation of problem (1.2).
An issue of main interest is providing a lower bound on the ground state energy (and thus on the spectrum) of H
V.
It is well-known that when V ≡ 0, then inf σ( H
V) = 0. On the other hand, if we take V ≤ 0, the kinetic energy ´
RN
|∇ ϕ |
2dx and the potential energy ´
RN
V ϕ
2dx are in competition in the quadratic form Q
Vand one could expect that
inf σ( H
V) < 0.
Actually, this depends on the potential V . For example, by recalling the Hardy inequality on R
N(for N ≥ 3)
N − 2 2
2ˆ
RN
ϕ
2| x |
2dx ≤ ˆ
RN
|∇ ϕ |
2dx, ϕ ∈ C
0∞( R
N\ { 0 } ), we get that if the potential V is such that
0 ≥ V ≥ −
N − 2 2
21
| x |
2,
then the spectrum of H
Vis still non-negative. This is an example of how Hardy-type inequalities
can be exploited in order to identify classes of negative potentials with non-negative spectrum.
1.2. Aim of the paper. In this paper we deal with a confined version of this problem. More precisely, we turn our attention to prescribed open sets Ω ⊂ R
N. We fix a potential V ∈ L
2loc(Ω) such that V ≤ 0 and consider the localized Schr¨ odinger operator with homogeneous boundary conditions H
Ω,V= − ∆ + V , this time acting on the domain
(1.3) D( H
Ω,V) := H
2(Ω) ∩ H
01(Ω) ∩ { u ∈ L
2(Ω) : V u ∈ L
2(Ω) } .
Here H
01(Ω) is the closure of C
0∞(Ω) in the Sobolev space H
1(Ω). This is still a symmetric and self-adjoint operator H
V: D( H
Ω,V) → L
2(Ω), with real spectrum σ( H
Ω,V). Observe that the hypothesis V ∈ L
2loc(Ω) entails as before the inclusion
C
0∞(Ω) ⊂ D( H
Ω,V),
thus the operator is densely defined. We define the associated quadratic form Q
Ω,V(ϕ) =
ˆ
Ω
|∇ ϕ |
2dx + ˆ
Ω
V ϕ
2dx, ϕ ∈ D( H
Ω,V).
The stationary equation (1.1) now reads
(1.4)
H
Ω,Vu = λ u in Ω, u = 0, in R
N\ Ω.
Equation (1.4) can be formally considered as a peculiar form of (1.1), where the potential V has the trapping property V = + ∞ in R
N\ Ω. This models the physical situation where the particle is
“trapped” in the confining region Ω.
The issue we tackle is the following
“find explicit pointwise bounds on the potential V assuring that the ground state energy of H
Ω,Vstays positive ”
In the vein of the example discussed above using Hardy’s inequality in the entire space, we will approach this problem by proving localized Hardy-type inequalities with suitable weights. A typical instance of these inequalities occurs when we limit ourselves to consider functions supported in a proper open subset Ω ⊂ R
Nand we use the distance d
Ω(x) := dist(x, ∂Ω) as a weight. In other words, one has
1 C
ˆ
Ω
ϕ
2d
2Ωdx ≤
ˆ
Ω
|∇ ϕ |
2dx, ϕ ∈ C
0∞(Ω).
However, the existence of such a constant C > 0 typically requires some conditions on the geometry
of the set Ω or on the regularity of its boundary. In this paper on the contrary, we will prove
alternative Hardy-type inequalities, with weights depending on solutions of peculiar elliptic partial
differential equations.
Roughly speaking, we will consider the solution w
q,Ωto the Lane-Emden equation
1with 1 ≤ q < 2 (1.5)
− ∆u = u
q−1in Ω, u = 0, in R
N\ Ω, u > 0, in Ω,
prove a Hardy inequality with weight depending on w
q,Ωand show that the condition 0 ≥ V & −
∇ w
q,Ωw
q,Ω2
, a. e. in Ω, leads to positivity of the spectrum of the Schr¨ odinger operator H
Ω,V.
The function w
q,Ωwill be called the Lane-Emden q − density of Ω, we refer to Definitions 2.5 and 2.8 below.
1.3. Main results. Let us now try to be more precise about our results. We first need to fix some definitions. For γ ≥ 1, we denote
(1.6) λ
2,γ(Ω) = inf
ϕ∈C0∞(Ω)
ˆ
Ω
|∇ ϕ |
2dx : k ϕ k
Lγ(Ω)= 1
. Henceforth we shall often work with the following class of sets.
Definition 1.1. We say that Ω ⊂ R
Nis an open set with positive spectrum if it is open and (1.7) λ
1(Ω) := λ
2,2(Ω) = inf
ϕ∈C0∞(Ω)
ˆ
Ω
|∇ ϕ |
2dx : k ϕ k
L2(Ω)= 1
> 0.
The main result of the paper is the following lower bound on the ground state energy of H
Ω,V. We refer to Theorem 6.2 and Corollary 6.3 for its proof.
Theorem 1.2. Let Ω ⊂ R
Nbe an open set with positive spectrum, and let V ∈ L
2loc(Ω). For an exponent 1 ≤ q < 2, we assume that
0 ≥ V ≥ − 1 4
∇ w
q,Ωw
q,Ω2
, a. e. in Ω.
Then the spectrum σ( H
Ω,V) of H
Ω,Vis positive and we have that inf σ( H
Ω,V) = inf
ϕ∈C0∞(Ω)
Q
Ω,V(ϕ) : ˆ
Ω
ϕ
2dx = 1
≥ 1
C λ
1(Ω), where C = C(N, q) > 0 is an explicit dimensional constant.
As stated above, the main tool we use to prove this result is an Hardy-type inequality, in which a weight involving the solution w
q,Ωof the Lane-Emden equation (1.5) enters. This is the content of the next result. For questions related to optimal choices of weights in Hardy-type inequalities, see [10] and the references therein.
1The terminology comes from astrophysics, where the Lane-Emden equation is 1
%2 d d%
%2du d%
+uγ= 0,
for a radially symmetric functionu:R3→R. The positive numberγis usually calledpolytropic index. Observe that for a radial functionudefined inR3, this is equivalent to
−∆u=uγ.
Though our paper is not concerned with astrophysics, we found useful to give a name to the equation and its solution.
Theorem 1.3 (Hardy-Lane-Emden inequality). Let 1 ≤ q < 2 and let Ω ⊂ R
Nbe an open set with positive spectrum. Then for every u ∈ C
0∞(Ω) and δ > 0 we have that
1 δ
1 − 1
δ ˆ
Ω
∇ w
q,Ωw
q,Ω2
ϕ
2dx + 1 δ
ˆ
Ω
ϕ
2w
2−qq,Ωdx ≤
ˆ
Ω
|∇ ϕ |
2dx.
We refer to Remark 3.2 for some comments about the proof of this result.
1.4. Plan of the paper. The paper is organized as follows: in Section 2, we define the Lane- Emden q − density of a set Ω ⊂ R
N, first under the assumption that Ω is bounded and then for a general open set. Then in Section 3 we prove the Hardy-Lane-Emden inequality of Theorem 1.3 for bounded open sets.
In Section 4 we show how the summability properties of the Lane-Emden densities are equivalent to the embedding of D
01,2(Ω) into suitable Lebesgue spaces. This part generalizes some results contained in the recent paper [3], by replacing the torsion function with any Lane-Emden q − density.
Though this section may appear unrelated to ground state energy estimates for H
Ω,V, some of its outcomes are used to extend (in Section 5) the Hardy-Lane-Emden inequality to open sets with positive spectrum.
The proof of Theorem 1.2 is then contained in Section 6, while Section 7 contains some appli- cations of our main result to some particular geometries (a ball, an infinite slab and a rectilinear wave-guide with circular cross-section).
We conclude the paper with an Appendix, containing a local L
∞estimate for subsolutions of the Lane-Emden equation, which is necessary in order to get the explicit lower bound on the ground state energy of H
Ω,V.
Acknowledgments. The first author would like to thank Douglas Lundholm for a discussion on Hardy inequalities and the so-called Ground State Representation in February 2017, during a visit to the Department of Mathematics of KTH (Stockholm). He also wishes to thank Erik Lindgren for the kind invitation. Remark 4.4 comes from an informal discussion with Guido De Philippis in December 2015, we wish to thank him.
The authors are members of the Gruppo Nazionale per l’Analisi Matematica, la Probabilit` a e le loro Applicazioni (GNAMPA) of the Istituto Nazionale di Alta Matematica (INdAM).
2. Preliminaries
2.1. Notation. Let Ω ⊂ R
Nbe an open set and define the norm on C
0∞(Ω) k ϕ k
D1,20 (Ω)
= ˆ
Ω
|∇ ϕ |
2dx
12
, ϕ ∈ C
0∞(Ω).
We consider the homogeneous Sobolev space D
01,2(Ω), obtained as the completion of C
0∞(Ω) with respect to the norm k · k
D1,20 (Ω). For N ≥ 3 this is always a functional space, thanks to Sobolev inequality but in dimension N = 1 or N = 2, this may fail to be even a space of distributions if Ω is “too big”, see for example [9, Remark 4.1].
Remark 2.1. For an open set with positive spectrum Ω, we have automatically continuity of the embedding D
01,2(Ω) , → L
2(Ω). Thus in this case D
1,20(Ω) is a functional space. Moreover, we have that
D
01,2(Ω) = H
01(Ω),
thanks to the fact that in this case ˆ
Ω
|∇ ϕ |
2dx
12and
ˆ
Ω
|∇ ϕ |
2dx
12+ ˆ
Ω
ϕ
2dx
12, are equivalent norms on C
0∞(Ω).
2.2. Lane-Emden densities: bounded sets. We start with the following auxiliary result.
Lemma 2.2. Let Ω ⊂ R
Nbe an open bounded set. For 1 ≤ q < 2, the variational problem
(2.1) min
ϕ∈D01,2(Ω)
1 2
ˆ
Ω
|∇ ϕ |
2dx − 1 q
ˆ
Ω
ϕ
qdx : ϕ ≥ 0 a. e. in Ω
,
admits a unique solution.
Proof. Since the absolute value of every minimizer of the functional ϕ 7→ 1
2 ˆ
Ω
|∇ ϕ |
2dx − 1 q
ˆ
Ω
| ϕ |
qdx, is also a minimizer of (2.1), problem (2.1) is equivalent to
min
ϕ∈D01,2(Ω)
1 2
ˆ
Ω
|∇ ϕ |
2dx − 1 q
ˆ
Ω
| ϕ |
qdx
.
The existence of a solution follows then by the Direct Methods in the Calculus of Variations, since the embedding D
01,2(Ω) , → L
q(Ω) is compact and D
1,20(Ω) is weakly closed.
As for uniqueness, we first suppose that Ω is connected. We observe that for q = 1 problem (2.1) is strictly convex, thus the solution is unique. For 1 < q < 2, we can use a trick by Brezis and Oswald based on the so-called Picone’s inequality, see [6, Theorem 1]. We reproduce their argument here for completeness. We first observe that a minimizer is a positive solution of the Lane-Emden equation
(2.2) − ∆u = u
q−1, in Ω,
with homogeneous Dirichlet boundary conditions. More precisely, for every ϕ ∈ D
1,20(Ω) it holds (2.3)
ˆ
Ω
h∇ u, ∇ ϕ i dx = ˆ
Ω
u
q−1ϕ dx.
We now suppose that (2.1) admits two minimizers u
1, u
2∈ D
1,20(Ω). By the minimum principle for superharmonic functions, u
1> 0 and u
2> 0 on Ω. Moreover, by standard Elliptic Regularity, u
1, u
2∈ L
∞(Ω). We fix ε > 0, then we test equation (2.3) for u
1with
ϕ = u
22u
1+ ε − u
1, and equation (2.3) for u
2with
ϕ = u
21u
2+ ε − u
2.
Summing up, we get that ˆ
Ω
∇ u
1, ∇ u
22u
1+ ε
dx − ˆ
Ω
|∇ u
1|
2dx + ˆ
Ω
∇ u
2, ∇ u
21u
2+ ε
dx − ˆ
Ω
|∇ u
2|
2dx
= ˆ
Ω
u
q−11u
1+ ε u
22dx − ˆ
Ω
u
q1dx + ˆ
Ω
u
q−12u
2+ ε u
21dx − ˆ
Ω
u
q2dx.
We now recall that (2.4)
∇ u, ∇ v
2u
≤ |∇ v |
2,
for v and u > 0 differentiable. This is precisely Picone’s inequality, see for example [2]. By observing that ∇ u
i= ∇ (u
i+ ε) and using (2.4) in the identity above, we conclude that
ˆ
Ω
u
q−11u
1+ ε u
22dx − ˆ
Ω
u
q1dx + ˆ
Ω
u
q−12u
2+ ε u
21dx − ˆ
Ω
u
q2dx ≤ 0.
We now take the limit as ε goes to 0. By Fatou’s Lemma, we obtain that ˆ
Ω
u
q−21u
22dx − ˆ
Ω
u
q1dx + ˆ
Ω
u
q−22u
21dx − ˆ
Ω
u
q2dx ≤ 0.
The previous terms can be recast into inequality ˆ
Ω
(u
22− u
21) (u
q−22− u
q−21) dx ≥ 0.
By using the fact the function t 7→ t
q−2is monotone decreasing, we get that u
1= u
2as desired.
Finally, if Ω is not connected, it is sufficient to observe that a solution of (2.1) must minimize the same functional on every connected component, due to the locality of the functional; since the solution is unique on every connected component, we get the conclusion in this case as well.
Remark 2.3 (About uniqueness). Uniqueness of the solution to (2.1) can also be inferred directly at the level of the minimization problem. It is sufficient to observe that the functional to be minimized is convex along curves of the form
γ
t=
(1 − t) ϕ
q0+ t ϕ
q11q
, t ∈ [0, 1], ϕ
0, ϕ
1∈ D
01,2(Ω) positive,
see [2, Proposition 2.6]. Then one can reproduce the uniqueness proof of [1]. For a different proof of the uniqueness for (2.2), we also refer to [11, Corollary 4.2].
Remark 2.4. It is useful to keep in mind that if u ∈ D
1,20(Ω) solves equation
− ∆u = t u
q−1, in Ω, for some t > 0, then the new function
v
t= t
q−21u, solves (2.2).
Definition 2.5. Let Ω ⊂ R
Nbe an open bounded set. For 1 ≤ q < 2, we define the Lane-Emden
q − density of Ω as the unique solution of (2.1). We denote such a solution by w
q,Ω. In the case
q = 1, we simply write w
Ωand call it torsion function of Ω.
The variational problem defining w
q,Ωis related to the optimal Poincar´e constant λ
2,q(Ω) defined in (1.6). This is the content of the next result, that we record for completeness. We omit the proof since it is based on a straightforward scaling argument.
Lemma 2.6. Let 1 ≤ q < 2 and let Ω ⊂ R be an open bounded set. Then we have
(2.5) min
ϕ∈D01,2(Ω)
1 2
ˆ
Ω
|∇ ϕ |
2dx − 1 q
ˆ
Ω
ϕ
qdx : ϕ ≥ 0 in Ω
= q − 2 2 q
1 λ
2,q(Ω)
2−qqand
(2.6)
ˆ
Ω
| w
q,Ω|
qdx
2−qq= 1
λ
2,q(Ω) .
2.3. Lane-Emden densities: general sets. We now want to define the Lane-Emden densities for a general open set, where the variational problem
inf
ϕ∈D1,20 (Ω)
1 2
ˆ
Ω
|∇ ϕ |
2dx − 1 q
ˆ
Ω
ϕ
qdx : ϕ ≥ 0 in Ω
, may fail to admit a solution.
We start with a sort of comparison principle for Lane-Emden densities.
Lemma 2.7. Let 1 ≤ q < 2 and let Ω
1⊂ Ω
2⊂ R
Nbe two open bounded sets. Then we have w
q,Ω1≤ w
q,Ω2.
Proof. We test the minimality of w
q,Ω1against ϕ = min { w
q,Ω1, w
q,Ω2} . After some simple manipu- lations, this gives
1 2
ˆ
{wq,Ω2<wq,Ω1}
|∇ w
q,Ω2|
2dx − 1 q
ˆ
{wq,Ω2<wq,Ω1}
w
qq,Ω2
dx
≥ 1 2
ˆ
{wq,Ω2<wq,Ω1}
|∇ w
q,Ω1|
2dx − 1 q
ˆ
{wq,Ω2<wq,Ω1}
w
q,Ωq1
dx.
We now add on both sides the term 1
2 ˆ
{wq,Ω2>wq,Ω1}
|∇ w
q,Ω2|
2dx − 1 q
ˆ
{wq,Ω2>wq,Ω1}
w
qq,Ω2
dx, thus if set U = max { w
q,Ω1, w
q,Ω2} , we get that
1 2
ˆ
Ω2
|∇ w
q,Ω2|
2dx − 1 q
ˆ
Ω2
w
q,Ωq 2dx ≥ 1 2
ˆ
Ω2
|∇ U |
2dx − 1 q
ˆ
Ω2
U
qdx.
By uniqueness of the minimizer w
q,Ω2, this gives U = w
q,Ω2. By recalling the definition of U, this
in turn yields the desired conclusion.
Thanks to the previous property, we can define the Lane-Emden density for every open set. In what follows, we set
Ω
R= Ω ∩ B
R(0), R > 0,
where B
R(0) is the N − dimensional open ball, with radius R and centered at the origin.
Definition 2.8. Let Ω ⊂ R
Nbe an open set. For 1 ≤ q < 2 we define w
q,Ω= lim
R→+∞
w
q,ΩR.
We observe that this definition is well-posed, since each w
q,ΩR∈ D
01,2(Ω
R) exists thanks to the boundedness of Ω
Rand the function
R 7→ w
q,ΩR(x), is monotone, thanks to Lemma 2.7.
Remark 2.9 (Consistency). When Ω ⊂ R
Nis an open bounded set or, more generally, is such that the embedding D
01,2(Ω) , → L
q(Ω) is compact, then the definition of w
q,Ωabove coincides with the variational one. For q = 1 this is proved in [3, Lemma 2.4], the other cases can be treated in exactly the same way. We skip the details.
3. Hardy-Lane-Emden inequalities
The following theorem, which is a generalization of [3, Theorem 4.3], is the main result of the present section. For simplicity, we state and prove the result just for open bounded sets, but it is easily seen that the same proof works for every open set Ω ⊂ R
Nsuch that the embedding D
01,2(Ω) , → L
q(Ω) is compact.
Theorem 3.1. Let 1 ≤ q < 2 and let Ω ⊂ R
Nbe an open bounded set. Then for every ϕ ∈ C
0∞(Ω) and δ > 0 we have
(3.1) 1
δ ˆ
Ω
"
1 − 1
δ
∇ w
q,Ωw
q,Ω2
+ 1
w
2−qq,Ω#
ϕ
2dx ≤ ˆ
Ω
|∇ ϕ |
2dx.
Proof. We recall that (3.2)
ˆ
Ω
h∇ w
q,Ω, ∇ ψ i dx = ˆ
Ω
w
q−1q,Ωψ dx,
for any ψ ∈ D
1,20(Ω). Let ϕ ∈ C
0∞(Ω) and let ε > 0, by taking in (3.2) the test function ψ = ϕ
2w
q,Ω+ ε , we get
(3.3)
ˆ
Ω
"
|∇ w
q,Ω|
2+ w
q−1q,Ω(w
q,Ω+ ε) (w
q,Ω+ ε)
2#
ϕ
2dx = 2 ˆ
Ω
ϕ
∇ w
q,Ω(w
q,Ω+ ε) , ∇ ϕ
dx.
By Young’s inequality, it holds ϕ
∇ w
q,Ω(w
q,Ω+ ε) , ∇ ϕ
≤ δ
2 |∇ ϕ |
2+ 1 2 δ
|∇ w
q,Ω|
2(w
q,Ω+ ε)
2ϕ
2for δ > 0. Thus we get
ˆ
Ω
"
|∇ w
q,Ω|
2+ w
q,Ωq−1(w
q,Ω+ ε) (w
q,Ω+ ε)
2#
ϕ
2dx ≤ δ ˆ
Ω
|∇ ϕ |
2dx + 1 δ
ˆ
Ω
|∇ w
q,Ω|
2(w
q,Ω+ ε)
2ϕ
2dx.
The previous inequality gives 1
δ ˆ
Ω
"
1 − 1
δ
|∇ w
q,Ω|
2(w
q,Ω+ ε)
2+ w
q,Ωq−1(w
q,Ω+ ε)
#
ϕ
2dx ≤ ˆ
Ω
|∇ ϕ |
2dx.
By recalling that ϕ is compactly supported in Ω and observing that
2∇ w
q,Ωw
q,Ω2
∈ L
1loc(Ω),
we conclude the proof by taking the limit as ε goes to 0 and appealing to the Monotone Convergence
Theorem.
Remark 3.2 (A comment on the proof). The idea of the previous proof comes from that of Moser’s logarithmic estimate for elliptic partial differential equations, see [14, page 586]. In regularity theory, this is an essential tool in order to establish the validity of Harnack’s inequality for solutions.
An alternative proof is based on Picone’s inequality (2.4). This goes as follows: one observes that the function W = w
1/δq,Ωlocally solves
− ∆W = − 1 δ w
1 δ−1
q,Ω
∆w
q,Ω− 1 δ
1 δ − 1
w
1 δ−2
q,Ω
|∇ w
q,Ω|
2= W
"
1
δ w
q−2q,Ω+ 1 δ
1 − 1
δ
∇ w
q,Ωw
q,Ω2
# . Thus we have
ˆ
Ω
"
1
δ w
q−2q,Ω+ 1 δ
1 − 1
δ
∇ w
q,Ωw
q,Ω2
#
W ψ dx = ˆ
Ω
h∇ W, ∇ ψ i dx,
for every ψ ∈ C
0∞(Ω). If we now take the test function ψ = ϕ
2/W and use inequality (2.4), we get the desired inequality.
This technique to obtain Hardy-type inequalities is sometimes referred to as Ground State Rep- resentation, see for example [12, Proposition 1].
As a consequence of the Hardy-Lane-Emden inequality, we record the following integrability properties of functions in D
1,20(Ω).
Corollary 3.3. Let 1 ≤ q < 2 and let Ω ⊂ R
Nbe an open bounded set. Then for every ϕ ∈ D
01,2(Ω) (3.4)
ˆ
Ω
∇ w
q,Ωw
q,Ω2
ϕ
2dx < + ∞ and
ˆ
Ω
ϕ
2w
2−qq,Ωdx < + ∞ . Moreover, if { ϕ
n}
n∈N⊂ D
01,2(Ω) converges strongly to ϕ ∈ D
01,2(Ω), then
n→∞
lim ˆ
Ω
∇ w
q,Ωw
q,Ω2
| ϕ
n− ϕ |
2dx = 0 and lim
n→∞
ˆ
Ω
| ϕ
n− ϕ |
2w
2−qq,Ωdx = 0.
2It is sufficient to remark that∇wq,Ω∈L2(Ω) and that by the strong minimum principle, we have wq,Ω≥cK>0 for everyKbΩ.
Proof. Let ϕ ∈ D
1,20(Ω), then there exists { ϕ
n}
n∈N⊂ C
0∞(Ω) converging to ϕ in D
01,2(Ω). By choosing δ = 2 in (3.1), we have that
1 4
ˆ
Ω
"
∇ w
q,Ωw
q,Ω2
+ 2
w
2−qq,Ω#
ϕ
2ndx ≤ ˆ
Ω
|∇ ϕ
n|
2dx.
By using the norm convergence in the right-hand side and Fatou’s Lemma in the left-hand side, we deduce the validity of (3.4) for ϕ.
In order to prove the second part of the statement, we observe that the first part of the proof also implies the validity of inequality (3.1) in D
1,20(Ω), for δ = 2. Plugging in ϕ
n− ϕ gives that
lim sup
n→∞
1 4
ˆ
Ω
"
∇ w
q,Ωw
q,Ω2
+ 2
w
2−qq,Ω#
| ϕ
n− ϕ |
2dx ≤ lim
n→∞
ˆ
Ω
|∇ ϕ
n− ∇ ϕ |
2dx = 0,
as desired.
As a consequence of Corollary 3.3 and thanks to the definition of D
01,2(Ω), we get the following Corollary 3.4. The Hardy-Lane-Emden inequality (3.1) is valid for every δ > 0 and u ∈ D
01,2(Ω).
4. Sobolev embeddings and densities
In this section, we consider general open sets and study the connections between the integrability of w
q,Ωand the embeddings of D
1,20(Ω) into Lebesgue spaces. For the case of the torsion function, i.e. when q = 1, related studies can be found in [3, 5, 7] and [8].
We start with a simple consequence of Theorem 3.1. This is valid for a general open set.
Lemma 4.1. Let Ω ⊂ R
Nbe an open set and 1 ≤ q < 2. Then for any ϕ ∈ C
0∞(Ω) it holds that ˆ
{x∈Ω :wq,Ω(x)<+∞}
ϕ
2w
2−qq,Ωdx ≤
ˆ
Ω
|∇ ϕ |
2dx.
Proof. Let B
R(0) be the ball of radius R centered in 0, we set Ω
R= Ω ∩ B
R(0) and w
R= w
q,ΩR. Let ϕ ∈ C
0∞(Ω), then for every R large enough the support of ϕ is contained in Ω
R. By using (3.1) on Ω
Rwith δ = 1, we get
ˆ
Ω
ϕ
2w
2−qRdx ≤
ˆ
Ω
|∇ ϕ |
2dx.
We conclude by letting R → + ∞ and by Fatou’s Lemma.
The following result is a generalization of [3, Theorem 1.2]. We point out that the equivalence between 1. and 2. below is a known fact in Sobolev spaces theory, see [13, Theorems 15.6.2].
Theorem 4.2. Let 1 ≤ q < 2 and let Ω ⊂ R
Nbe an open set. Then for every q ≤ γ < 2 the following three facts are equivalent
1. the embedding D
1,20(Ω) , → L
γ(Ω) is continuous;
2. the embedding D
1,20(Ω) , → L
γ(Ω) is compact;
3. w
q,Ω∈ L
2−q 2−γγ
(Ω).
Moreover, we have the double-sided estimates
(4.1) 1 ≤ λ
2,γ(Ω)
ˆ
Ω
w
2−q 2−γγ q,Ω
dx
2−γγ
≤ 2 − γ
γ − 2 (q − 1)
2 − q 2 − γ
2, where λ
2,γ(Ω) is the optimal Poincar´ e constant defined in (1.6).
Proof. As announced above, the equivalence 1. ⇐⇒ 2. is already known, see also [3, Theorem 1.2]
for a different proof. It is sufficient to prove the equivalence 1. ⇐⇒ 3.
Let us suppose that the embedding D
01,2(Ω) , → L
γ(Ω) is continuous. As always, we set B
R(0) the ball of radius R centered in 0, Ω
R= Ω ∩ B
R(0) and w
R= w
q,ΩR. Then by testing (2.3) with w
Rβfor some β ≥ 1, we get
ˆ
ΩR
w
β+q−1Rdx = β 2
β + 1
2ˆ
ΩR
∇ w
β+1 2
R
2
dx
≥ β 2
β + 1
2λ
2,γ(Ω
R) ˆ
ΩR
w
β+1 2 γ
R
dx
2γ≥ β 2
β + 1
2λ
2,γ(Ω) ˆ
ΩR
w
β+1 2 γ
R
dx
2γ. By choosing
3β = γ − 2 (q − 1) 2 − γ , from the previous estimate we get
ˆ
ΩR
w
2−q 2−γγ
R
dx
2−γγ≤ 2 − γ γ − 2 (q − 1)
2 − q 2 − γ
21 λ
2,γ(Ω)
By Fatou’s Lemma, we can take the limit as R goes to + ∞ and get the desired integrability of w
q,Ω, together with the upper estimate in (4.1).
Suppose now that w
q,Ω∈ L
2−q2−γγ(Ω), this implies that w
q,Ω< + ∞ almost everywhere in Ω. We take u ∈ C
0∞(Ω), then by H¨ older’s inequality and Lemma 4.1 we have
ˆ
Ω
| ϕ |
γdx = ˆ
Ω
| ϕ |
γw
(2−q)γ 2
q,Ω
w
(2−q)γ 2
q,Ω
dx ≤ ˆ
Ω
ϕ
2w
q,Ω2−qdx
!
γ2ˆ
Ω
w
2−q 2−γγ q,Ω
dx
2−γ2≤ ˆ
Ω
|∇ ϕ |
2dx
γ2ˆ
Ω
w
2−q 2−γγ q,Ω
dx
2−γ2.
We conclude by density of C
0∞(Ω) in D
1,20(Ω) that the embedding D
1,20(Ω) , → L
γ(Ω) is continuous.
Moreover, we also obtain the lower bound in (4.1).
The following result generalizes [3, Theorem 1.3] and [4, Theorem 9], by allowing any Lane- Emden densities in place of the torsion function.
3Observe thatβ≥1 thanks to the fact thatq≤γ <2.
Proposition 4.3. Let 1 ≤ q < 2 and let Ω ⊂ R
Nbe an open set. Then we have that λ
1(Ω) > 0 ⇐⇒ w
q,Ω∈ L
∞(Ω).
Moreover, we have that
(4.2) λ
1(Ω)
q−21≤ k w
q,Ωk
L∞(Ω)≤
2
NC
22 C
2−q+ 4
2−q1λ
1(Ω)
q−21, where C is the same constant appearing in (A.1).
Proof. We suppose that w
q,Ω∈ L
∞(Ω). This in particular implies that w
q,Ω< + ∞ almost every- where in Ω. Then for any ϕ ∈ C
0∞(Ω) we have that
ˆ
Ω
ϕ
2dx = ˆ
Ω
ϕ
2w
2−qq,Ωw
2−qq,Ωdx
≤ ˆ
Ω
ϕ
2w
2−qq,Ωdx
!
k w
q,Ωk
2−qL∞(Ω)≤ k w
q,Ωk
2−qL∞(Ω)ˆ
Ω
|∇ ϕ |
2dx, the last inequality being due to Lemma 4.1. This shows that
w
q,Ω∈ L
∞(Ω) for 1 ≤ q < 2 = ⇒ λ
1(Ω) > 0, together with the lower bound in (4.2).
The converse implication is more involved and we adapt the proof of [4, Theorem 9], which deals with the case q = 1. Without loss of generality we can suppose Ω to be bounded and smooth;
indeed, the general case can be then covered by considering a family of smooth bounded sets approaching Ω from inside.
For ease of notation we set w := w
q,Ωand we suppose that w(0) = k w k
L∞(Ω). This can be done up to translating Ω. Moreover we can extend w to 0 outside Ω. Since ∂Ω is regular, we get by means of Hopf’s Lemma that the extended function, which we still denote by w, satisfies
(4.3) − ∆w ≤ w
q−1,
in the weak sense. Let R > 0 to be fixed, and let ζ be a cut-off Lipschitz function such that 0 ≤ ζ ≤ 1, ζ = 1 in B
R(0), ζ = 0 in R
N\ B
2R(0), |∇ ζ | ≤ 1
R . From the variational characterization of λ
1(Ω), we have
(4.4) λ
1(Ω) ≤ ˆ
Ω
|∇ (w ζ) |
2dx ˆ
Ω
(w ζ)
2dx
= ˆ
Ω
|∇ w |
2ζ
2+ 2 w ζ h∇ w, ∇ ζ i + |∇ ζ |
2w
2dx ˆ
Ω
w
2ζ
2dx
.
By using the positive test function w ζ
2into the weak formulation of (4.3), we get that ˆ
Ω
|∇ w |
2ζ
2dx + 2 ˆ
Ω
w ζ h∇ w, ∇ ζ i dx = ˆ
Ω
h∇ w, ∇ (w ζ
2) i dx ≤ ˆ
Ω
w
qζ
2dx.
Thus, by recalling that w attains its maximum in 0 and using the properties of ζ, from (4.4) we obtain that
(4.5) λ
1(Ω) ≤ ˆ
Ω
|∇ ζ |
2w
2+ w
qζ
2dx ˆ
Ω
w
2ζ
2dx
≤ 2
Nω
Nw(0)
2ˆ R
N−2+ w(0)
qR
NBR(0)
w
2dx .
We use now the local L
∞− L
2estimate of Lemma A.1 to handle the denominator. Indeed, by (A.1) with α = 2 we have that
ˆ
BR(0)
w
2dx ≥ ω
NR
N1
C w(0) − R
2
2−q2!
2.
By spending this information in (4.5), we end up with
λ
1(Ω) ≤ 2
NR
−2w(0)
2+ w(0)
q1
C w(0) − R
2
2−q2!
2.
By choosing
R = 2
w(0) 2 C
(2−q)/2, we obtain the inequality
λ
1(Ω) ≤ 2
N4 C
2w(0)
2−q1 4
2 C
2−q+ 1
, and thus
w(0) ≤
2
NC
22 C
2−q+ 4
2−q1λ
1(Ω)
q−21.
This concludes the proof.
Remark 4.4 (Super-homogeneous embeddings). A closer inspection of the proof reveals that with exactly the same argument we can prove the following stronger statement: for every 1 ≤ q < 2 and 2 ≤ γ < 2
∗, we have that
(4.6) λ
2,γ(Ω) > 0 ⇐⇒ w
q,Ω∈ L
∞(Ω),
where
2
∗= 2 N
N − 2 , for N ≥ 3 and 2
∗= + ∞ , for N ∈ { 1, 2 } . Observe that (4.6) implies in particular that
(4.7) D
01,2(Ω) , → L
2(Ω) ⇐⇒ D
01,2(Ω) , → L
γ(Ω), for 2 < γ < 2
∗.
For the implication = ⇒ , it is sufficient to reproduce the proof above, using the variational charac- terization of λ
2,γ(Ω) and the L
∞estimate (A.1), this time with α = γ.
For the converse implication, it is sufficient to use the Gagliardo-Nirenberg interpolation inequal- ity (see for example [3, Proposition 2.6])
ˆ
RN
| u |
γdx
γ1≤ C ˆ
RN
| u |
2dx
1−ϑ2ˆ
RN
|∇ u |
2dx
ϑ2,
where C = C(N, γ) > 0 and
ϑ =
1 − 2 γ
N
2 , 2 < γ < 2
∗.
This shows that if D
01,2(Ω) , → L
2(Ω) is continuous, then D
1,20(Ω) , → L
γ(Ω) is continuous as well.
We leave the details to the interested reader.
We point out that the equivalence (4.7) can also be found in [13, Theorem 15.4.1]. The proof there is different.
We conclude this section with the following simple result which we record for completeness.
Proposition 4.5. Let 1 ≤ q < 2 and let Ω ⊂ R
Nbe an open set such that the embedding D
1,20(Ω) , → L
q(Ω) is continuous. Then the embedding D
1,20(Ω) , → L
2(Ω) is compact.
Proof. We already know by Theorem 4.2 that the continuity of the embedding D
1,20(Ω) , → L
q(Ω) is equivalent to its compactness. Then it is sufficient to use the Gagliardo-Nirenberg inequality
ˆ
RN
| u |
2dx
12
≤ C ˆ
RN
| u |
qdx
1−ϑq
ˆ
RN
|∇ u |
2dx
ϑ2
, where C = C(N, q) > 0 and
ϑ = 1 − q
2
2 N (2 − q) N + 2 q .
This guarantess that every bounded sequence { u
n}
n∈N⊂ D
1,20(Ω) strongly converging in L
q(Ω), strongly converges in L
2(Ω) as well. This gives the desired conclusion.
Remark 4.6. The converse implication of the previous proposition does not hold. Indeed, let { r
i}
i∈N⊂ R be a sequence of strictly positive numbers, such that
i→∞
lim r
i= 0 and
∞
X
i=0
r
2 2−γ+N
i
= + ∞ , for every 1 ≤ γ < 2.
For example, one could take r
i= 1/ log(2 + i). We then define the sequence of points { x
i}
i∈N⊂ R
Nby
x
0= (0, . . . , 0),
x
i+1= (r
i+ r
i+1, 0, . . . , 0) + x
i, and the disjoint union of balls
Ω =
∞
[
i=0