Citation: Federico Bernini, Simone Secchi. Existence of solutions for a perturbed problem with logarithmic potential in R2[J]. Mathematics in Engineering, 2020, 2(3): 438-458. doi: 10.3934/mine.2020020
[1] | Wenjing Liao, Mauro Maggioni, Stefano Vigogna . Multiscale regression on unknown manifolds. Mathematics in Engineering, 2022, 4(4): 1-25. doi: 10.3934/mine.2022028 |
[2] | Pier Domenico Lamberti, Michele Zaccaron . Spectral stability of the curlcurl operator via uniform Gaffney inequalities on perturbed electromagnetic cavities. Mathematics in Engineering, 2023, 5(1): 1-31. doi: 10.3934/mine.2023018 |
[3] | Emilia Blåsten, Fedi Zouari, Moez Louati, Mohamed S. Ghidaoui . Blockage detection in networks: The area reconstruction method. Mathematics in Engineering, 2019, 1(4): 849-880. doi: 10.3934/mine.2019.4.849 |
[4] | G. Gaeta, G. Pucacco . Near-resonances and detuning in classical and quantum mechanics. Mathematics in Engineering, 2023, 5(1): 1-44. doi: 10.3934/mine.2023005 |
[5] | Juan-Carlos Felipe-Navarro, Tomás Sanz-Perela . Semilinear integro-differential equations, Ⅱ: one-dimensional and saddle-shaped solutions to the Allen-Cahn equation. Mathematics in Engineering, 2021, 3(5): 1-36. doi: 10.3934/mine.2021037 |
[6] | Yoshikazu Giga, Hirotoshi Kuroda, Michał Łasica . The fourth-order total variation flow in Rn. Mathematics in Engineering, 2023, 5(6): 1-45. doi: 10.3934/mine.2023091 |
[7] | Antonio De Rosa, Luca Lussardi . On the anisotropic Kirchhoff-Plateau problem. Mathematics in Engineering, 2022, 4(2): 1-13. doi: 10.3934/mine.2022011 |
[8] | Rupert L. Frank, Tobias König, Hynek Kovařík . Energy asymptotics in the Brezis–Nirenberg problem: The higher-dimensional case. Mathematics in Engineering, 2020, 2(1): 119-140. doi: 10.3934/mine.2020007 |
[9] | Lucrezia Cossetti . Bounds on eigenvalues of perturbed Lamé operators with complex potentials. Mathematics in Engineering, 2022, 4(5): 1-29. doi: 10.3934/mine.2022037 |
[10] | Serena Federico, Gigliola Staffilani . Sharp Strichartz estimates for some variable coefficient Schrödinger operators on R×T2. Mathematics in Engineering, 2022, 4(4): 1-23. doi: 10.3934/mine.2022033 |
The Newton kernel ΦN:RN→R is defined by
ΦN(x)={Γ(N−22)4πN/2|x|N−2if N≥312πlog1|x|if N=2. |
The non-local partial differential equation
−Δu+au=[ΦN⋆|u|2]uin RN, | (1.1) |
where a is a positive function, was proposed in the study of quantum physics of electrons in a ionic crystal (the so-called Pekar polaron model) for N=3. The same equation can also be seen as a coupling of quantum physics with Newtonian gravitation: indeed, the system
{iψt−Δψ+E(x)ψ+γwψ=0Δw=|ψ|2 |
in the unknown ψ:RN×R→R, ψ=ψ(x,t), reduces to the single equation
−Δu+au+γ[ΦN⋆|u|2]u=0 |
via the ansatz ψ(x,t)=e−iλtu(x) with λ∈R and a(x)=E(x)+λ. E.H. Lieb proved in [11] that (1.1) possesses, in dimension N=3, a unique ground state solution which is positive and radially symmetric. E. Lenzmann proved in [10] that this solution is also non-degenerate. The analysis of (1.1) in dimension N=3 is heavily based on the algebraic properties of the kernel Φ3, in particular its homogeneity. Lieb's proof of existence carries over to N=4 and N=5, while no solution with finite energy can exist in dimension N≥6, see [6].
In this note we consider (1.1) in the plane, i.e., when N=2. The appearance of the logarithm in Φ2 changes drastically the setting of the problem, which has been an open field of study for several years. One of the main obstructions to a straightforward analysis in the planar case is the lack of positivity of the kernel Φ2.
Some preliminary numerical results contained in [9] encouraged Ph. Choquard, J. Stubbe and M. Vuffray to prove the existence of a unique positive radially symmetric solution by applying a shooting method, see [7]. But only in very recent years have variational methods been used to solve (1.1) for N=2: the formal definition of a Euler functional
I(u)=12∫R2(|∇u|2+a|u|2)+18π∫R2×R2log|x−y||u(x)|2|u(y)|2dxdy |
is not consistent with the metric structure of the Sobolev space H1(R2).
Stubbe proposed in [14] a variational setting for (1.1) in dimension two within the closed subspace
X={u∈H1(R2)∣∫R2log(1+|x|)|u(x)|2dx<+∞} |
endowed with the norm
‖u‖2X=∫R2(|∇u|2+a|u|2)+∫R2log(1+|x|)|u(x)|2dx. |
Although this space permits to use variational methods, several difficulties arise from the logarithmic term.
Using this functional approach, S. Cingolani and T. Weth (see [8]) proved some existence results for (1.1) under either a periodicity assumption on the potential a, or the action of a suitable group of transformations. Uniqueness and monotonicity of positive solutions are also proved.
Later on, D. Bonheure, S. Cingolani and J. Van Schaftingen (see [6]) proved that the positive solution u of (1.1) with a>0 is non-degenerate, in the sense that the only solutions of the linearized equation associated to (1.1) are the (linear combinations of) the two partial derivatives of u.
Motivated by these results, we consider the following perturbed equation, based on (1.1):
−Δu+au−12π[log1|⋅|⋆u2]u=εh(x)|u|p−1uin R2. | (1.2) |
The quantity ε plays the rôle of a "small" perturbation, and the function h is a "weight" for the local nonlinearity |u|p−1u. We refer to the next Sections for the precise assumptions we make.
We will face the problem of constructing solutions to (1.2) by means of a general technique in Critical Point Theory, introduced by A. Ambrosetti and M. Badiale in [1,2,3]. We refer to [4] for a presentation in book form. For the reader's convenience, we summarize here the main ideas of this method.
Suppose we are given a (real) Hilbert space X and a functional Iε∈C2(X) of the form
Iε(#)=I0(#)+εG(#). |
Here I0∈C2(X) is the so-called unperturbed functional, while ε∈R is a (small) perturbation parameter. We will suppose that there exists a (smooth) manifold Z of dimension d<∞, such that every z∈Z is a critical point of I0.
Letting W=(TzZ)⊥ for z∈Z, we look for solutions to the equation I′ε(u)=0 of the form u=z+w, where z∈Z and w∈W. We can split the equation I′ε(u)=0 into two equations by means of the orthogonal projection P:X→W:
{PI′ε(z+w)=0(I−P)I′ε(z+w)=0. | (1.3) |
We will assume that the following conditions hold:
(ND) for all z∈Z, we have TzZ=kerI″0(z);
(Fr) for all z∈Z, we have that the linear operator I″0(z) is Fredholm with index zero.
Remark 1.1. The condition (ND) can be seen as a non-degeneracy assumption, since it is always true that TzZ⊂kerI″0(z), by definition of Z.
It is possible to show that the first equation of system (1.3) can be (uniquely) solved with respect to w=w(ε,z), with z∈Z and ε sufficiently small. The main result of this perturbation technique can be summarized in the following statement.
Theorem 1.2 ([4]). Suppose that the function Φε:Z→R defined by Φε(z)=Iε(z+w(ε,z)) possesses, for |ε| sufficiently small, a critical point zε∈Z. Then uε=zε+w(ε,zε) is a critical point of Iε=I0+εG.
As it should be clear, the perturbation method of Ambrosetti and Badiale leans on the effect of the function h, which breaks the invariance of I0 under translations. As such, the existence of a critical point of the function Φε depends crucially on the behavior of h.
We split our existence results into two categories. The first one assumes that the weight function h is not only bounded, but also sufficiently integrable over R2; because of this, we can consider this results as a local existence result.
Theorem 1.3. Let p>1 and h∈L∞(R2)∩Lq(R2) for some q>1. Moreover, suppose that
(h1) ∫R2h(x)|z0|p+1dx≠0.
Then Eq. (1.2) has a solution provided |ε| is small enough.
It is possibile to drop the integrability condition on h, at the cost of a more delicate analysis of the implicit function w=w(ε,z) that describes Φε. We have the following global result.
Theorem 1.4. Let p>2 and suppose that h satisfies
(h2) h∈L∞(R2) and lim|x|→∞h(x)=0.
Then for all |ε| small, Eq. (1.2) has a solution.
We highlight that our results differ from those appearing in the literature for several reasons. First of all, the non-degeneracy property appearing in Proposition 2.5 can be used as a basis for further investigation. Moreover, the right-hand side of Eq. (1.2) may (and indeed must) depend on x; no symmetry requirement, like radial symmetry, is needed in our proofs.
The paper is organized as follows. In Section 2 we first give the precise assumptions for our problem, then we recall some known results (classical and not) and we prove some properties for energy functional, such as regularity. In Section 3 we present the proof of the main theorems.
Consider the equation
−Δu+au−12π[log1|⋅|⋆u2]u=εh(x)|u|p−1uin R2, | (2.1) |
with a>0, h∈L∞(R2) and p>1. We introduce the function space
X={u∈H1(R2)∣∫R2|u(x)|2log(1+|x|)dx<∞}, |
endowed with the norm
‖u‖2X=‖u‖2H1(R2)+|u|2∗, |
where
‖u‖2H1(R2)=∫R2(|∇u(x)|2+a|u(x)|2)dx|u|2∗=∫R2|u(x)|2log(1+|x|)dx. |
The norm ‖⋅‖X is associated naturally to an inner product. Let
Iε(u)=I0(u)+εG(u) | (2.2) |
be the energy functional associated to the equation, where
I0(u)=12∫R2(|∇u(x)|2+au2(x))dx−18π∫R2×R2log1|x−y||u(x)|2|u(y)|2dxdy |
and
G(u)=−1p+1∫R2h(x)|u(x)|p+1dx. |
We define (see [8,14]) the symmetric bilinear forms
B1(u,v)=∫R2×R2log(1+|x−y|)u(x)v(y)dxdy,B2(u,v)=∫R2×R2log(1+1|x−y|)u(x)v(y)dxdy, |
and
B(u,v)=B1(u,v)−B2(u,v)=∫R2×R2log(|x−y|)u(x)v(y)dxdy, | (2.3) |
since for all r>0 we have
logr=log(1+r)−log(1+1r). |
Remark 2.1. The definitions above are restricted to measurable functions u, v:R2→R such that the corresponding double integral is well defined in the Lebesgue sense.
In order to find estimates for B1 and B2 we recall a classical result of Measure Theory.
Theorem 2.2 (Hardy-Littlewood-Sobolev's inequality [12]). Let p>1, q>1 and 0<λ<N with 1p+λN+1q=2. If f∈Lp(RN) and g∈Lq(RN), then there exists a sharp constant C(N,λ,p), independent of f and g, such that
∫RN×RN|f(x)g(y)||x−y|λdxdy≤C(N,λ,p)‖f‖Lp(RN)‖g‖Lq(RN). | (2.4) |
The sharp constant satisfies
C(N,λ,p)≤N(N−λ)(|SN−1|N)λN1pq((λ/N1−1p)λN+(λ/N1−1q)λN). |
If p=q=2N2N−λ, then
C(N,λ,p)=C(N,λ)=πλ2Γ(N/2−λ/2)Γ(N−λ/2)(Γ(N/2)Γ(N))−1+λ/N. |
In this case there is equality in (2.4) if and only if g≡cf with c constant and
f(x)=A(γ2+|x−α|2)−(2N−λ)/2 |
for some A∈R, γ∈R∖{0} and α∈RN.
We note that, since
log(1+|x−y|)≤log(1+|x|+|y|)≤log(1+|x|)+log(1+|y|), |
we have by Schwarz's inequality
|B1(uv,wz)|≤∫R2×R2[log(1+|x|)+log(1+|y|)]|u(x)v(x)||w(y)z(y)|dxdy≤|u|∗|v|∗‖w‖L2(R2)‖z‖L2(R2)+‖u‖L2(R2)‖v‖L2(R2)|w|∗|z|∗ | (2.5) |
for u, v, w, z∈X. Next, since 0≤log(1+r)≤r for all r>0, we have by Hardy-Littlewood-Sobolev's inequality
|B2(u,v)|≤∫R2×R21|x−y|u(x)v(y)dxdy≤C‖u‖L43(R2)‖v‖L43(R2), | (2.6) |
for u,v∈L43(R2), for some constant C>0. In particular, from (2.5) we have
B1(u2,u2)≤2|u|2∗‖u‖2L2(R2) | (2.7) |
for all u∈X and from (2.6) we have
B2(u2,u2)≤C‖u‖4L83(R2) | (2.8) |
for all u∈L83(R2).
Proposition 2.3. The functional Iε is of class C2(X).
Proof. The proof is similar to [8,Lemma 2.2], so we just sketch the main ideas. Recalling (2.7), (2.8), the assumption h∈L∞(R2) and the fact that X is compactly embedded into Ls(R2), s∈[2,+∞) (see [5,6,8] for a proof) we have
|Iε(u)|≤12‖u‖H1(R2)+18πB(u2,u2)+1p+1ε‖h‖∞‖u‖p+1Lp+1(R2)≤12‖u‖H1(R2)+14π|u|2∗‖u‖2L2(R2)+18πC‖u‖4L43(R2)+‖h‖∞‖u‖p+1X<+∞. |
The first Gâteaux derivative of Iε along v is
I′ε(u)v=∫R2(∇u⋅∇v+auv)dx−12π∫R2×R2log1|x−y|u2(x)u(y)v(y)dxdy−ε∫R2h(x)|u|p−1uvdx. | (2.9) |
We add and subtract ∫R2u(x)v(x)log(1+|x|)dx to recover the scalar product of X, so we obtain
I′ε(u)v=(u|v)X−12πB(u2,uv)−∫R2u(x)v(x)log(1+|x|)dx−ε∫R2h(x)|u|p−1uvdx. |
Now,
|I′ε(u)v|≤(u|v)X+12πB(u2,uv)+∫R2u(x)v(x)log(1+|x|)dx+ε∫R2|h(x)||u|p|v|dx≤(u|v)X+12π|u|2∗‖uv‖L2(R2)+12π‖u‖2L2(R2)|uv|∗+12πC‖u‖2L83(R2)‖uv‖L43(R2)+ε‖h‖∞‖u‖ppp′‖v‖p |
The second Gâteaux derivative of Iε(u) along (v,w) is
I″ε(u)(v,w)=∫R2(∇v⋅∇w+avw)dx−12π∫R2×R2log1|x−y|u2(x)v(y)w(y)dxdy−1π∫R2×R2log1|x−y|u(x)v(x)u(y)w(y)dxdy−(p−1)ε∫R2h(x)|u|p−1wdx−ε∫R2h(x)|u|p−1vwdx. |
In this case, we add and subtract ∫R2v(x)w(x)log(1+|x|)dx, hence
I″ε(u)(v,w)=(v|w)X−∫R2v(x)w(x)log(1+|x|)dx−12πB(u2,vw)−1πB(uv,uw)−(p−1)ε∫R2h(x)|u|p−1wdx−ε∫R2h(x)|u|p−1vwdx. |
Finally,
|I″ε(u)(v,w)|≤(v|w)X+∫R2v(x)w(x)log(1+|x|)dx+12πB(u2,vw)+1πB(uv,uw)+(p−1)ε∫R2|h(x)||u|p−1|w|dx+ε∫R2|h(x)||u|p−1|v||w|dx≤(v|w)X+∫R2v(x)w(x)log(1+|x|)dx+12π|u|2∗‖vw‖L2(R2)+12π‖u‖2L2(R2)|v|∗|w|∗+12πC‖u‖2L83(R2)‖vw‖L43(R2)+1π|u|∗|v|∗‖u‖L2(R2)‖w‖L2(R2)+1π‖u‖L2(R2)‖v‖L2(R2)|u|∗|w|∗+1πC‖u‖L2(R2)‖v‖L2(R2)‖u‖L2(R2)‖w‖L2(R2)+(p−1)ε‖h‖∞‖u‖p−1(p−1)p′‖w‖p+ε‖h‖∞‖u‖p−1(p−1)p′‖vw‖p. |
It is now standard to conclude that the first and the second Gâteaux derivatives are continuous (with respect to u∈X), so that Iε∈C2(X).
Critical points of the unperturbed functional I′0(u)=0 are solutions of the equation
−Δu+au−12π[log1|⋅|⋆u2]u=0in R2, | (2.10) |
which admits for every a>0 a unique — up to translations — radially symmetric solution u∈X (see [8], Theorem 1.3).
Since (2.10) is invariant under translations, we can consider zξ(x)=u(x−ξ), with ξ∈R2. The manifold
Z={zξ∣ξ∈R2} |
is therefore a critical manifold for I0. We want to show that the manifold Z satisfies the properties (ND) and (Fr).
The (ND) property follows from the following theorem, proved in [6].
Theorem 2.4. If a>0 and u∈X is a radial solution of (2.10) then there exists μ∈(0,∞) such that
u(x)=(μ+o(1))√|x|(log|x|)1/4exp(−√Me−aM∫eaM|x|1√(logsds) |
with
M=12π∫R2|u|2. |
To prove that (Fr) holds, we actually show that I″0(zξ) is a compact perturbation of the identity operator.
Proposition 2.5. I″0(zξ)=L−K, where L is a continuous invertible operator and K is a continuous linear compact operator in X.
Proof. We recall that
I″0(zξ)(v,w)=∫R2(∇v⋅∇w+avw)dx−12π∫R2×R2log1|x−y|z2ξ(x)v(y)w(y)dxdy−1π∫R2×R2log1|x−y|zξ(x)v(x)zξ(y)w(y)dxdy |
We add and subtract ∫R2log(1+|x|)v(x)w(x)dx, hence
I″0(zξ)(v,w)=∫R2(∇v⋅∇w+avw)dx+∫R2log(1+|x|)v(x)w(x)dx−∫R2log(1+|x|)v(x)w(x)dx−12π∫R2×R2log1|x−y|z2ξ(x)v(y)w(y)dxdy−1π∫R2×R2log1|x−y|zξ(x)v(x)zξ(y)w(y)dxdy, |
and so
I″0(zξ)(v,w)=(v|w)X−∫R2log(1+|x|)v(x)w(x)dx−12π∫R2×R2log1|x−y|z2ξ(x)v(y)w(y)dxdy−1π∫R2×R2log1|x−y|zξ(x)v(x)zξ(y)w(y)dxdy. |
The second derivative is therefore the linear operator defined by
L(zξ):φ↦−Δφ+(a−w)φ+2zξ(log2π⋆(zξφ)), |
where
w(x)=−12π∫R2log|x−y||zξ(y)|2dy,x∈R2. |
Following the proof in [13], Lemma 15, let {vn}n, {wn}n be two sequences in X such that ‖vn‖≤1, ‖wn‖≤1, vn⇀v0 and wn⇀w0. Without loss of generality we can assume v0=w0=0, so that
vn⇀0,wn⇀0. |
From the compact embedding of X in Ls(R2) for s≥2, we can say that
vn→0,wn→0 | (2.11) |
in Ls(R2), for every s≥2. We compute, using Theorem 2.4,
∫R2×R2log1|x−y|zξ(x)vn(x)zξ(y)wn(y)dxdy≤C‖zξvn‖L43(R2)‖zξwn‖L43(R2). |
Now, by Hölder's inequality,
lim supn→+∞‖zξvn‖43L43(R2)≤lim supn→+∞(∫R2|zξ(x)|4dx)13(∫R2|vn(x)|2dx)23=0 |
thanks to Theorem 2.4 and (2.11).
Similarly,
lim supn→∞|∫R2×R2log1|x−y|zξ(x)vn(x)zξwn(y)dxdy|=0. | (2.12) |
This proves that the linear operator
φ↦2zξ(log2π⋆(zξφ)) |
is compact. At this point, we should notice that the linear operator
φ↦−Δφ+(a−w)φ |
is not invertible on X. To overcome this difficulty, we set
c2=12π∫R2|zξ(x)|2dx |
and rewrite L(zξ) as follows:
L(zξ):φ↦−Δφ+(a+c2log(1+|x|))φ−(c2log(1+|x|)+w)φ+2zξ(log2π⋆(zξφ)). |
Since
lim|x|→+∞(w(x)+c2log(1+|x|))=0 |
by [8,Proposition 2.3], the multiplication operator
φ↦(c2log(1+|x|)+w)φ |
is compact. We may conclude that the functional I″0(zξ) is of the form L−K where
Lφ=−Δφ+(a+c2log(1+|x|))φ |
is a linear, continuous, invertible operator and K is a linear, continuous, compact operator. A different proof, based on a direct computation, appears in [?].
Since the properties (ND) and (Fr) hold we can say that, for |ε| small, the reduced functional has the following form:
Φε(zξ)=Iε(zξ+wε(ξ))=c0+εG(zξ)+o(ε), |
with c0=I0(zξ).
Let Γ:R2→R be the function defined as
Γ(ξ)=G(zξ)=−1p+1∫R2h(x)|zξ|p+1dx,ξ∈R2. |
Lemma 3.1. Suppose that h∈L∞(R2)∩Lq(R2) for some q>1. Then
lim|ξ|→∞Γ(ξ)=0. |
Proof. By the Hölder inequality,
|Γ(ξ)|≤1p+1∫R2|h(x)||zξ|p+1dx≤1p+1(∫R2|h(x)|qdx)1q(∫R2|zξ|(p+1)q′dx)1q′ |
and since zξ decays to zero as |ξ|→∞ we have
|Γ(ξ)|≤C(∫R2|zξ|(p+1)q′dx)1q′→0as |ξ|→+∞. |
The proof is completed.
Thanks to the previous Lemma we can now prove the existence of local solutions for (2.1).
Proof of Theorem 1.3. The hypothesis of the previous Lemma are satisfied, so we have that Γ(ξ) goes to 0 as |ξ|→∞. From (h1) follows that Γ(0)=−1p+1∫R2h(x)|z0|p+1dx≠0. Then Γ is not identically zero and follows that Γ has a maximum or a minimum on R2 and the existence of a solution follows from Theorem 2.16 in [4].
As before, we call P=Pξ:X→Wξ the orthogonal projection onto Wξ=(TzξZ)⊥, ~Wξ:=⟨zξ⟩⊕(TzξZ) and Rξ(w)=I′0(zξ+w)−I″0(zξ)[w].
Remark 3.2. By the variational characterization of the Mountain-Pass solution u as in [4,Remark 4.2], the spectrum of PI″0(u) has exactly one negative simple eigenvalue with eigenspace spanned by u itself. Moreover, λ=0 is an eigenvalue with multiplicity N and eigenspace spanned by Diu, i=1,…,N and there exists κ>0 such that
(PI″0(u)[v]|v)≥κ‖v‖2,∀v⊥⟨u⟩⊕TuZ, |
and hence the rest of the spectrum is positive.
We prove the following
Theorem 3.3. (i) There is C>0 such that ‖(PI″0(zξ))−1‖L(Wξ,Wξ)≤C, for every ξ∈R2,
(ii) Rξ(w)=o(‖w‖), uniformly with respect to ξ∈R2.
Proof. Since zξ is a Mountain-Pass solution, Remark 3.2 holds, hence it suffices to show that there exists κ>0 such that
(PI″0(zξ)[v]∣v)≥κ‖v‖2,∀ξ∈R2,∀v⊥~Wξ. |
For any fixed ξ∈R2, say ξ=0, PI″0(z0)=PI″0(u) is invertible and there exists κ>0 such that
(PI″0(u)[v]∣v)≥κ‖v‖2,∀v∈˜W:=⟨u⟩⊕TuZ. |
Let vξ(x)=v(x+ξ), then
(PI″0(zξ)[v]∣v)=P[∫R2(|∇v|2+av2)dx−12π∫R2×R2log1|x−y|z2ξ(x)v2(y)dxdy−1π∫R2×R2log1|x−y|zξ(x)v(x)zξ(y)v(y)dxdy] |
and thanks to the change of variables x=t+ξ and y=s+ξ we obtain
(PI″0(zξ)[v]∣v)=P[∫R2(|∇v(t+ξ)|2+av2(t+ξ))dt−12π∫R2×R2log1|t−s|z2ξ(t+ξ)v2(s+xi)dtds−1π∫R2×R2log1|t−s|zξ(t+ξ)v(t+ξ)zξ(s+ξ)v(s+ξ)dtds]=P[∫R2(|∇vξ|2+a(vξ)2)dt−12π∫R2×R2log1|x−y|z20(t)|vξ(s)|2dtds−1π∫R2×R2log1|x−y|z0(t)vξ(t)z0(s)vξ(s)dtds]=PI″0(u)(vξ,vξ). |
Moreover, vξ⊥˜W whenever v⊥~Wξ, hence
(PI″0(zξ)[(v,v)]v]∣v)=(PI″0(u)[vξ]∣vξ)≥κ‖vξ‖2=κ‖v‖2,∀ξ∈R2,∀v⊥~Wξ |
and so (i) si true.
To prove (ii) we observe that
Rξ(w)=I′0(zξ+w)(v)−I″0(zξ)(w,v)=∫R2[∇(zξ+w)⋅∇v+a(zξ+w)v]dx−12π∫R2×R2log1|x−y|(zξ+w)2(x)(zξ+w)(y)v(y)dxdy−∫R2(∇w⋅∇v+awv)dx+12π∫R2×R2log1|x−y|z2ξ(x)w(y)v(y)dxdy+1π∫R2×R2log1|x−y|zξ(x)w(x)zξ(y)v(y)dxdy. |
After some computations we obtain
Rξ(w)=∫R2(∇zξ⋅∇v+azξv)dx−12π∫R2×R2log1|x−y|z2ξ(x)zξ(y)v(y)dxdy−12π∫R2×R2log1|x−y|z2ξ(x)w(y)v(y)dxdy−1π∫R2×R2log1|x−y|zξ(x)w(x)zξ(y)v(y)dxdy−1π∫R2×R2log1|x−y|zξ(x)w(x)w(y)v(y)dxdy−12π∫R2×R2log1|x−y|w2(x)zξ(y)v(y)dxdy−12π∫R2×R2log1|x−y|w2(x)w(y)v(y)dxdy+12π∫R2×R2log1|x−y|z2ξ(x)w(y)v(y)dxdy+1π∫R2×R2log1|x−y|zξ(x)w(x)zξ(y)v(y)dxdy |
and since zξ is critical point we finally have
Rξ(w)=−1π∫R2×R2log1|x−y|zξ(x)w(x)w(y)v(y)dxdy−12π∫R2×R2log1|x−y|w2(x)zξ(y)v(y)dxdy−12π∫R2×R2log1|x−y|w2(x)w(y)v(y)dxdy. |
By (2.7), (2.8), (2.4), Hölder's inequality we have
|Rξ(w)|≤1π(|zξ|∗|w|∗‖w‖L2‖v‖L2+‖zξ‖L2‖w‖L2|w|∗|v|∗+C1‖zξ‖L43‖w‖2L43‖v‖L43)+12π(|w|2∗‖zξ‖L2‖v‖L2+‖w‖2L2|zξ|∗|v|∗+C2‖w‖2L83‖zξ‖L43‖v‖L43)+12π(|w|2∗‖w‖L2‖v‖L2+‖w‖2L2|w|∗|v|∗+C3‖w‖2L83‖w‖L43‖v‖L43), |
by the compact embedding of X into Ls(R2), s≥2
|Rξ(w)|≤1π(‖zξ‖X‖w‖2X‖v‖X+‖zξ‖X‖w‖2X‖v‖X+C1‖zξ‖X‖w‖2X‖v‖X)+12π(‖w‖2X‖zξ‖X‖v‖X+‖w‖2X‖zξ‖X‖v‖X+C2‖w‖2X‖zξ‖X‖v‖X)+12π(‖w‖3X‖v‖X+‖w‖3X‖v‖X+C3‖w‖3X‖v‖X) | (3.1) |
and finally
|Rξ(w)|≤C4‖zξ‖X‖w‖2X‖v‖X+C5‖w‖2X‖zξ‖X‖v‖X+C6‖w‖3X‖v‖X. |
Hence,
|Rξ(w)|‖w‖X≤C4‖zξ‖X‖w‖X‖v‖X+C5‖w‖X‖zξ‖X‖v‖X+C6‖w‖2X‖v‖X |
and this goes to 0 as ‖w‖X→0 uniformly with respect to ξ∈R2.
This Lemma allows us to use Lemma 2.21 in [4], so there exists ε0>0 such that for all |ε|≤ε0 and all ξ∈R2 the auxiliary equation PI′ε(zξ+w)=0 has a unique solution wε(zξ) with
limε→0‖wε(zξ)‖=0, | (3.2) |
uniformly with respect to ξ∈R2.
We now prove
Lemma 3.4. There exists ε1>0 such that for all |ε|≤ε1, the following result holds:
lim|ξ|→∞wξ=0,strongly in X. |
Proof. We first show two preliminaries results:
(a) wξ weakly converges in X to some wε,∞∈X as |ξ|→∞. Moreover, the weak limit wε,∞ is a weak solution of
−Δwε,∞+awε,∞−12π[log1|⋅|⋆w2ε,∞]wε,∞=εh(x)|wε,∞|p−1wε,∞; | (3.3) |
(b) wε,∞=0.
As a consequence of (3.2) we have that wε(zξ) weakly converges in X to some wε,∞∈X, as |ξ|→∞. Recall that wε(zξ) is a solution of the auxialiary equation PI′ε(zξ+wξ(zξ))=0, namely
−Δwε,ξ+awε,ξ−12π[log1|⋅|⋆w2ε,ξ]wε,ξ−12π[log1|⋅|⋆z2ξ]wε,ξ−1π[log1|⋅|⋆z2ξ](zξ+wε,ξ)−12π[log1|⋅|⋆w2ε,ξ]zξ=εh(x)|zξ+wε,ξ|p−1(zξ+wε,ξ)−zp−1ξ−2∑i=1aiDizξ, |
where
ai=∫R2(εh(x)|zξ+wε,ξ|p−1−zp−1ξ)Dizξdx, |
and Di denotes the partial derivative with respect to xi.
Let v be any test function, then
∫R2(∇wε,ξ⋅∇v+awε,ξv)dx−12π∫R2×R2log1|x−y|w2ε,ξ(x)wε,ξ(y)v(y)dxdy=12π∫R2×R2log1|x−y|z2ξ(x)wε,ξ(y)v(y)dxdy+1π∫R2×R2log1|x−y|z2ξ(x)(zξ+wε,ξ)(y)v(y)dxdy+12π∫R2×R2log1|x−y|w2ε,ξ(x)zξ(y)v(y)dxdy+∫R2εh(x)|zξ+wε,ξ|p−1v(x)dx−∫R2zp−1ξ(x)v(x)dx−2∑i=1ai∫R2Dizξ(x)v(x)dx. |
Following the computations in the proof of Proposition 2.5, in particular (2.12), and by Theorem 2.4 we can say that
12π∫R2×R2log1|x−y|z2ξ(x)wε,ξ(y)v(y)dxdy→0,1π∫R2×R2log1|x−y|z2ξ(x)(zξ+wε,ξ)(y)v(y)dxdy→0,12π∫R2×R2log1|x−y|w2ε,ξ(x)zξ(y)v(y)dxdy→0, |
as |ξ|→∞.
Now, we need to pass to the limit in
∫R2εh(x)|zξ+wε,ξ|p−1v(x)dx. |
In order to do that, we show that
lim|ξ|→0∫R2zp−1−kξwkε,ξvdx=0,∀k∈[0,p−1). | (3.4) |
We split the integral as
∫R2zp−1−kξwkε,ξvdx=∫|x|≤ρzp−1−kξwkε,ξvdx+∫|x|>ρzp−1−kξwkε,ξvdx |
where ρ>0. Using Hölder's inequality with p=k+1, so p′=pp−1−k, we obtain
|∫|x|≤ρzp−1−kξwkε,ξvdx|≤(∫|x|≤ρ|z(p−1−k)p′ξdx)1p′(∫|x|≤ρ|wε,ξ|kp|v|pdx)1p≤C(∫|x|≤ρ|zξ|pdx)1p′ |
and this goes to 0 as ρ goes to ∞. On the other hand,
|∫|x|>ρzp−1−kξwkε,ξvdx|≤(∫|x|>ρ|z(p−1−k)p′ξ|wε,ξ|kp′dx)1p′(∫|x|>ρ|v|pdx)1p, |
and since v is a test function this integral goes to 0 as ρ goes to ∞.
Hence, (3.4) holds and then
∫R2εh(x)|zξ+wε,ξ|p−1v(x)dx→∫R2εh(x)|wε,∞|p−1v(x)dx,. |
Moreover,
∫R2|wε,ξ|p−1v(x)dx→∫R2|w∞|p−1v(x)dx,∫R2εh(x)|wε,ξ|p−1v(x)dx→∫R2εh(x)|w∞|p−1v(x)dx, |
as |ξ|→∞ and again by Theorem 2.4,
∫R2zp−1ξ(x)v(x)dx→0,∫R2Dizξ(x)v(x)dx→0, |
as |ξ|→∞.
Finally, we obtain
∫R2(∇wε,∞⋅∇v+awε,∞v)dx−12π∫R2×R2log1|x−y|w2ε,∞(x)wε,∞(y)v(y)dxdy=∫R2εh(x)|wε,∞|p−1v(x)dx, |
thus wε,∞ is a weak solution of (3.3), namely (a) holds.
By (3.2) we have that lim|ε|→0wε,∞=0. Since the unique solution w∈X of
−Δw+aw−12π[log1|⋅|⋆w2]w=εh(x)|w|p−1w |
with small norm is w=0. To show that, we need to prove that there exists a constant C>0 such that ‖w‖X≥C.
Consider the first Gâteaux derivative, computed in (2.9), evaluated at w along w, namely
0=I′ε(w)w=∫R2(|∇w|2+aw2)dx−12π∫R2×R2log1|x−y|w2(x)w2(y)dxdy−ε∫R2h(x)|w|p+1dx, |
thus
‖w‖2H1(R2)=12π∫R2×R2log1|x−y|w2(x)w2(y)dxdy+ε∫R2h(x)|w|p+1dx, |
and thus
‖w‖2H1(R2)≤12π∫R2×R2log1|x−y|w2(x)w2(y)dxdy+ε∫R2|h(x)||w|p+1dx. | (3.5) |
Now, observe that from (2.3) we have
−∫R2×R2log1|x−y||w(x)|2|w(y)|2dxdy=∫R2×R2log(1+|x−y|)|w(x)|2|w(y)|2dxdy−∫R2×R2log(1+1|x−y|)w2(x)w2(y)dxdy, |
hence
∫R2×R2log(1+1|x−y|)w2(x)w2(y)dxdy=∫R2×R2log(1+|x−y|)w2(x)w2(y)dxdy+∫R2×R2log(1|x−y|)w2(x)w2(y)dxdy |
and finally
∫R2×R2log(1+1|x−y|)w2(x)w2(y)dxdy≤∫R2×R2log(1|x−y|)w2(x)w2(y)dxdy. |
Moreover,
ε∫R2|h(x)||w|p+1dx≤C1‖w‖p+1Lp+1(R2)≤C2‖w‖p+1H1(R2). |
By (3.5) we obtain
‖w‖2H1(R2)≤∫R2×R2log(1|x−y|)w2(x)w2(y)dxdy+ε∫R2|h(x)||w|p+1dx≤C‖w‖4H1(R2)+C2‖w‖p+1H1(R2)≤C3‖w‖ηH1(R2), |
where η=max{p+1,4}>2 since p>1; recalling that ‖w‖η−2H1(R2)≤C4‖w‖η−2X we obtain
‖w‖η−2X≥1C5>0. |
From this we infer that wε,∞=0, provided that |ε|≪1. Hence (b) is true.
We recall that wε,ξ satisfies
wε,ξ=(PI″0(zξ))−1[εPG′(zξ+wε,ξ)−PRξ(wε,ξ)], | (3.6) |
where
G′(zξ+wε,ξ)=−∫R2h(x)|zξ+wε,ξ|p−1(zξ+wε,ξ)dx |
and
Rξ(wε,ξ)=−1π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)wε,ξ(y)v(y)dxdy−12π∫R2×R2log1|x−y|w2ε,ξ(x)zξ(y)v(y)dxdy−12π∫R2×R2log1|x−y|w2ε,ξ(x)wε,ξ(y)v(y)dxdy. |
From (3.6) and Theorem 3.3 it follows
‖wε,ξ‖2≤C[|ε||(G′(zξ+wε,ξ|wε,ξ))|+|(Rξ(wε,ξ)|wε,ξ)|]. | (3.7) |
We infer that
|(G′(zξ+wε,ξ|wε,ξ))|≤∫R2|h(x)||zξ+wε,ξ|p|wε,ξ(x)|dx≤‖h‖∞(∫R2|zξ|p|wε,ξ|dx+∫R2|wε,ξ|p+1dx) |
and this goes to 0 as |ξ|→∞ by the compact embedding of X in Ls(R2) for all s≥2, (a) and (b) proved above.
Then, by (3.1) we find that
|(Rξ(wε,ξ)|wε,ξ)|≤≤1π(‖zξ‖X‖wε,ξ‖2X‖wε,ξ‖X+‖zξ‖X‖wε,ξ‖2X‖wε,ξ‖X+C1‖zξ‖X‖wε,ξ‖2X‖wε,ξ‖X)+12π(‖wε,ξ‖2X‖zξ‖X‖wε,ξ‖X+‖wε,ξ‖2X‖zξ‖X‖wε,ξ‖X+C2‖wε,ξ‖2X‖zξ‖X‖wε,ξ‖X)+12π(‖wε,ξ‖3X‖wε,ξ‖X+‖wε,ξ‖3X‖wε,ξ‖X+C3‖wε,ξ‖3X‖wε,ξ‖X) |
thus,
|(Rξ(wε,ξ)|wε,ξ)|≤C4‖wε,ξ‖4X+o(ε). |
Inserting the above inequality in (3.7) we obtain that
‖wε,ξ‖2X≤C4‖wε,ξ‖4X+o(ε), as |ξ|→∞ |
and passing to the limit we find that
lim|ξ|→∞‖wε,ξ‖2X≤C4lim|ξ|→∞‖wε,ξ‖4X. |
Since wε,ξ is small in X as |ε|→0 we conclude that
lim|ξ|→∞‖wε,ξ‖X=0 provided |ε|≪1. |
Now we are ready to prove the existence of a global solution for the problem (2.1).
Proof of Theorem 1.4. Consider the reduced function Φε(ξ)=Iε(zξ+wε,ξ), namely
Φε(ξ)=12‖zξ+wε,ξ‖2H1(R2)−18π∫R2×R2log1|x−y|(zξ+wε,ξ)2(x)(zξ+wε,ξ)2(y)dxdy−εp+1∫R2h(x)|zξ+wε,ξ|p+1dx. | (3.8) |
From I0(zξ)=12‖zξ‖2H1(R2)−18π∫R2×R2log1|x−y|z2ξ(x)z2ξ(y)dxdy and setting c0=I0(zξ) we have that
12‖zξ‖2H1(R2)=c0+18π∫R2×R2log1|x−y|z2ξ(x)z2ξ(y)dxdy. |
Moreover, −Δzξ+azξ−12π[log1|⋅|∗z2ξ]zξ=0 implies
(zξ|wε,ξ)H1(R2)=12π∫R2×R2log1|x−y|z2ξ(x)zξ(y)wε,ξ(y)dxdy. |
Substituting into (3.8) we obtain
Φε(ξ)=12‖zξ‖2H1(R2)+12‖wε,ξ‖2H1(R2)+(zξ|wε,ξ)H1(R2)−18π∫R2×R2log1|x−y|(zξ+wε,ξ)2(x)(zξ+wε,ξ)2(y)dxdy−1p+1∫R2h(x)|zξ+wε,ξ|p+1dx=c0+18π∫R2×R2log1|x−y|z2ξ(x)z2ξ(y)dxdy+12‖wε,ξ‖2H1(R2)+12π∫R2×R2log1|x−y|z2ξ(x)zξ(y)wε,ξ(y)dxdy−18π∫R2×R2log1|x−y|(zξ+wε,ξ)2(x)(zξ+wε,ξ)2(y)dxdy−εp+1∫R2h(x)|zξ+wε,ξ|p+1dx, |
thus
Φε(ξ)=c0+18π∫R2×R2log1|x−y|z2ξ(x)z2ξ(y)dxdy+12‖wε,ξ‖2H1(R2)+12π∫R2×R2log1|x−y|z2ξ(x)zξ(y)wε,ξ(y)dxdy−18π∫R2×R2log1|x−y|z2ξ(x)z2ξ(y)dxdy−14π∫R2×R2log1|x−y|z2ξ(x)zξ(y)wε,ξ(y)dxdy−18π∫R2×R2log1|x−y|z2ξ(x)w2ε,ξ(y)dxdy−14π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)z2ξ(y)dxdy−12π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)zξ(y)wε,ξ(y)dxdy−14π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)w2ε,ξ(y)dxdy−18π∫R2×R2log1|x−y|w2ε,ξ(x)z2ξ(y)dxdy−14π∫R2×R2log1|x−y|w2ε,ξ(x)zξ(y)wε,ξ(y)dxdy−18π∫R2×R2log1|x−y|w2ε,ξ(x)w2ε,ξ(y)dxdy−εp+1∫R2h(x)|zξ+wε,ξ|p+1dx, |
and after a short computation
Φε(ξ)=c0+12‖wε,ξ‖2H1(R2)+12π∫R2×R2log1|x−y|z2ξ(x)zξ(y)wε,ξ(y)dxdy−14π∫R2×R2log1|x−y|z2ξ(x)zξ(y)wε,ξ(y)dxdy−18π∫R2×R2log1|x−y|z2ξ(x)w2ε,ξ(y)dxdy−14π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)z2ξ(y)dxdy−12π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)zξ(y)wε,ξ(y)dxdy−14π∫R2×R2log1|x−y|zξ(x)wε,ξ(x)w2ε,ξ(y)dxdy−18π∫R2×R2log1|x−y|w2ε,ξ(x)z2ξ(y)dxdy−14π∫R2×R2log1|x−y|w2ε,ξ(x)zξ(y)wε,ξ(y)dxdy−18π∫R2×R2log1|x−y|w2ε,ξ(x)w2ε,ξ(y)dxdy−εp+1∫R2h(x)|zξ+wε,ξ|p+1dx. |
Now, by repating the same arguments used in Proposition 2.5 all the integrals over R2×R2 converge to 0.
Moreover, by Lemma 3.4 we have ‖wε,ξ‖2H1(R2)→0 as |ξ|→+∞ and by Minkowski's inequality, Proposition 2.4, Lemma 3.4 and (h2):
lim|ξ|→+∞∫R2h(x)|zξ+wε,ξ|p+1dx=0. |
Hence,
lim|ξ|→+∞Φε(ξ)=c0. |
This means that either Φε is constant, or it has a maximum or minimum. In any case Φε has a critical point and we can apply Theorem 2.23 in [4] to find a solution for problem (2.1).
The authors are grateful to the anonymous referee for their comments and suggestions about the first version of the paper.
The authors declare no conflict of interest.
[1] | Ambrosetti A, Badiale M (1998) Homoclinics: Poincaré-Melnikov type results via a variational approach. Ann Inst H Poincaré Anal Non Linéaire 15: 233-252. |
[2] |
Ambrosetti A, Badiale M (1998) Variational perturbative methods and bifurcation of bound states from the essential spectrum. Proc Roy Soc Edinburgh Sect A 128: 1131-1161. doi: 10.1017/S0308210500027268
![]() |
[3] |
Ambrosetti A, Badiale M, Cingolani S (1997) Semiclassical states of nonlinear Schrödinger equations. Arch Ration Mech Anal 140: 285-300. doi: 10.1007/s002050050067
![]() |
[4] | Ambrosetti A, Malchiodi A (2006) Perturbation methods and semilinear elliptic problems on Rn, Basel: Birkhäuser Verlag. |
[5] | Bernini F, Mugnai D (2020) On a logarithmic Hartree equation. Adv Nonlinear Anal 9: 850-865. |
[6] |
Bonheure D, Cingolani S, Van Schaftingen J (2017) The logarithmic Choquard equation: Sharp asymptotics and nondegeneracy of the groundstate. J Funct Anal 272: 5255-5281. doi: 10.1016/j.jfa.2017.02.026
![]() |
[7] | Choquard P, Stubbe J, Vuffray M (2008). Stationary solutions of the Schrödinger-Newton model-an ODE approach. Differ Integral Equ 21: 665-679. |
[8] | Cingolani S, Weth T (2016) On the planar Schrödinger-Poisson system. Ann Inst H Poincaré Anal Non Linéaire 33: 169-197. |
[9] |
Harrison R, Moroz I, Tod KP (2003) A numerical study of the Schrödinger-Newton equations. Nonlinearity 16: 101-122. doi: 10.1088/0951-7715/16/1/307
![]() |
[10] |
Lenzmann E (2009) Uniqueness of ground states for pseudorelativistic Hartree equations. Anal PDE 2: 1-27. doi: 10.2140/apde.2009.2.1
![]() |
[11] |
Lieb EH (1977) Existence and uniqueness of the minimizing solution of Choquard's nonlinear equation. Stud in Appl Math 57: 93-105. doi: 10.1002/sapm197757293
![]() |
[12] | Lieb EH, Loss M (2001) Analysis, 2 Eds., Vol 14 of Graduate Studies in Mathematics, Providence: American Mathematical Society. |
[13] |
Secchi S (2010) A note on Schrödinger-Newton systems with decaying electric potential. Nonlinear Anal 72: 3842-3856. doi: 10.1016/j.na.2010.01.021
![]() |
[14] | Stubbe J (2008) Bound states of two-dimensional Schrödinger-Newton equations. arXiv:0807.4059. |