betaFeedback

WorksMathematics

One bubble against a ring: new sign-changing solutions of the Brezis-Nirenberg problem

Abstract

We consider the Brezis–Nirenberg problem −Δu=∣u∣2∗−2u+εu-Δu=|u|^{2^*-2}u+\varepsilon u in the unit ball of RN\mathbb R^N, N≥5N\ge5, with Dirichlet boundary conditions, and we look for sign-changing solutions which, as ε→0+\varepsilon\to0^+, blow up positively at the centre and negatively at the kk vertices of a regular polygon. The key observation is that, for this configuration, the gradient of the reduced energy factorises, so that the whole finite-dimensional problem is governed by the zeros of a single explicit function of the radius of the polygon. Its behaviour depends on the dimension: when N=5N=5 such solutions exist for every k≥2k\ge2, and the ring of negative peaks approaches the boundary at distance O(1/k)O(1/k); when N≥6N\ge6 there are two solutions for 2≤k≤62\le k\le6, while no solution of this type exists for kk large. For 2≤k≤62\le k\le6 the nodal set of every such solution reaches the boundary of the ball, in contrast with what happens for the almost critical problem. Nevertheless, the two solutions found for N≥6N\ge6 have very different nodal structures: numerically, for 3≤k≤63\le k\le6 one of them has only two nodal domains, and its nodal set meets the plane of the peaks along a closed curve around the centre, while the other one can have as many as k+1k+1 nodal domains. These results extend to higher dimensions a configuration previously found in dimension four.

Introduction and main results

In this paper we continue the study started in [18] of sign-changing blowing-up solutions of the Brezis–Nirenberg problem

{−Δu=∣u∣2∗−2u+εuin Ω,u=0on ∂Ω,(1)\begin{cases} -\Delta u=|u|^{2^*-2}u+\varepsilon u & \text{in }\Omega,\\ u=0 & \text{on }\partial\Omega, \tag*{(1)} \end{cases}

where Ω⊂RN\Omega\subset\mathbb{R}^N, N≥5N\geq5, is a smooth bounded domain, ε>0\varepsilon>0 is a small parameter and, as usual, $2^*=\frac{2N}{N-2}$ is the critical Sobolev exponent. Problem (1) was introduced by Brezis and Nirenberg in the celebrated paper [5], where it is shown that, if N≥4N\geq4, it has a positive solution for every ε∈(0,λ1(Ω))\varepsilon\in(0,\lambda_1(\Omega)), λ1(Ω)\lambda_1(\Omega) being the first eigenvalue of −Δ-\Delta in Ω\Omega with Dirichlet boundary conditions, while in dimension N=3N=3 positive solutions exist only for ε\varepsilon in a proper subinterval of (0,λ1(Ω))(0,\lambda_1(\Omega)). On the other hand, by the Pohozaev identity, (1) has no nontrivial solutions for ε≤0\varepsilon\leq0 when Ω\Omega is star-shaped. The lack of compactness of the embedding H01(Ω)↪L2∗(Ω)H^1_0(\Omega)\hookrightarrow L^{2^*}(\Omega) is reflected in the behaviour of the solutions as ε→0+\varepsilon\to0^+: they may blow up, and their profile near each blow-up point is described by a bubble, i.e. an extremal of the Sobolev inequality [2, 23]; the bubbles are the only positive solutions of the limit problem in $\mathbb{R}^N [6].

For positive solutions this phenomenon is by now well understood. Han [9] and Rey [22] proved that minimal energy solutions blow up, as ε→0\varepsilon\to0, at a critical point of the Robin function.

2020 Mathematics Subject Classification. Primary 35J61; Secondary 35B33, 35B44, 35J25, 35B05.

Key words and phrases. Brezis–Nirenberg problem; critical Sobolev exponent; sign-changing solutions; blowing-up solutions; multi-peak solutions; nodal set; Lyapunov–Schmidt reduction.

G. M. Rago and G. Vaira acknowledge support from the INDAM-GNAMPA project “Singolarità, nonlinearità e interazioni non locali: nuovi approcci analitici per equazioni differenziali e sistemi” (CUP E53C25002010001). of Ω\Omega, and Rey [22] showed that, conversely, in dimension N≥5N \ge5 non-degenerate critical points of the Robin function generate families of solutions blowing up at them. Positive solutions concentrating at several points were constructed by Musso and Pistoia [17]: in this case the role of the Robin function is played by a finite-dimensional function which involves both the Robin function and the Green function of Ω\Omega, and the geometry of the domain becomes crucial. A fine description of multi-bubble blow-up has been obtained by König and Laurain [12, 13] (see also Druet [7] for the three-dimensional case), and the construction of bubbling solutions in general bounded domains in low dimensions has been recently addressed in [14, 15].

The picture for sign-changing solutions is much richer and still far from complete. Sign-changing solutions may blow up at different points with different signs [4, 11, 16], or at the same point, as a tower of bubbles with alternating signs [10, 20]; nodal radial solutions in the ball, for which the dependence on the dimension is particularly evident, have been studied in [1, 19], and related phenomena for critical problems with a Hardy potential in [8]. In all these results the dimension plays a crucial role, since the concentration rate of the bubbles, and hence the structure of the finite-dimensional problem obtained by the reduction, depends on NN: the scales are of order ε1N−4\varepsilon^{\frac{1}{N-4}} when N≥5N \ge5, while they are exponentially small in 1/ε1/\varepsilon when N=4N = 4.

In dimension N≥5N \ge5, sign-changing multi-bubble solutions of critical and supercritical problems were studied by Micheletti and Pistoia [16]. For the Brezis–Nirenberg problem their reduction leads to a finite-dimensional function which depends simultaneously on the concentration points and on the individual scales. To describe it, and to explain the difficulties we have to face, let us introduce some notation which will be used throughout the paper. Let G(x,y)G(x,y) be the Green function of −Δ-\Delta in Ω\Omega with Dirichlet boundary conditions and let HH be its regular part, i.e.

G(x,y)=1ωN∣x−y∣N−2−H(x,y).G(x,y)=\frac{1}{\omega_N\lvert x-y\rvert^{N-2}}-H(x,y).

where ωN:=(N−2)meas⁡(SN−1)\omega_N := (N-2)\operatorname{meas}(S^{N-1}), SN−1S^{N-1} is the unit sphere of RN\mathbb{R}^N and, for every y∈Ωy \in\Omega,

−ΔxH(x,y)=0in Ω,H(x,y)=1ωN∣x−y∣N−2on ∂Ω.-\Delta_x H(x,y)=0 \quad\text{in }\Omega,\qquad H(x,y)=\frac{1}{\omega_N\lvert x-y\rvert^{N-2}}\quad\text{on }\partial\Omega.

The leading term of the regular part, τ(x):=H(x,x)\tau(x):=H(x,x), is called the Robin function of Ω\Omega. Given an integer k≥2k \ge2, we set

Ωk∗:={ξ=(ξ1,…,ξk)∈Ωk:ξi≠ξj for i≠j}\Omega_k^*:=\left\{\xi=(\xi_1,\ldots,\xi_k)\in\Omega^k:\xi_i\ne\xi_j\text{ for }i\ne j\right\}

and, for ξ∈Ωk∗\xi\in\Omega_k^* and σ=(σ1,…,σk)∈{−1,1}k\sigma=(\sigma_1,\ldots,\sigma_k)\in\{-1,1\}^k, we define the symmetric matrix

Mσ(ξ)=(τ(ξ1)−σ1σ2G(ξ1,ξ2)⋯−σ1σkG(ξ1,ξk)−σ1σ2G(ξ1,ξ2)τ(ξ2)⋯−σ2σkG(ξ2,ξk)⋮⋮⋱⋮−σ1σkG(ξ1,ξk)−σ2σkG(ξ2,ξk)⋯τ(ξk)).M^\sigma(\xi)= \begin{pmatrix} \tau(\xi_1) & -\sigma_1\sigma_2G(\xi_1,\xi_2) & \cdots& -\sigma_1\sigma_kG(\xi_1,\xi_k)\\ -\sigma_1\sigma_2G(\xi_1,\xi_2) & \tau(\xi_2) & \cdots& -\sigma_2\sigma_kG(\xi_2,\xi_k)\\ \vdots& \vdots& \ddots& \vdots\\ -\sigma_1\sigma_kG(\xi_1,\xi_k) & -\sigma_2\sigma_kG(\xi_2,\xi_k) & \cdots& \tau(\xi_k) \end{pmatrix}.

Mσ(ξ)M^\sigma(\xi) is the Green–Robin interaction matrix: its diagonal entries describe the interaction of each bubble with the boundary, while the off-diagonal entries describe the interaction between distinct bubbles. In our notation, the leading part of the reduced energy in [16] is

Fkσ(d,ξ)=aN2[∑j=1kτ(ξj)djN−2−∑i,j=1i≠jkσiσjG(ξi,ξj)diN−22djN−22]−bN∑j=1kdj2(2)F_k^\sigma(d,\xi)=\frac{a_N}{2}\left[\sum_{j=1}^{k}\tau(\xi_j)d_j^{N-2}-\sum_{\substack{i,j=1\\i\ne j}}^{k}\sigma_i\sigma_jG(\xi_i,\xi_j)d_i^{\frac{N-2}{2}}d_j^{\frac{N-2}{2}}\right]-b_N\sum_{j=1}^{k}d_j^2 \tag*{(2)}

or equivalently, setting κj:=djN−22\kappa_j := d_j^{\frac{N-2}{2}},

Fkσ(κ,ξ)=aN2⟨Mσ(ξ)κ,κ⟩−bN∑j=1kκj4N−2,(3)F_k^\sigma(\boldsymbol{\kappa},\boldsymbol{\xi}) = \frac{a_N}{2}\langle M^\sigma(\boldsymbol{\xi})\boldsymbol{\kappa},\boldsymbol{\kappa}\rangle- b_N\sum_{j=1}^{k}\kappa_j^{\frac{4}{N-2}}, \tag*{(3)}

where aN,bN>0a_N,b_N>0 are given in (8). Thus the geometry of the domain and the nodal pattern enter the reduced problem simultaneously.

The general reduction allows one to prescribe arbitrary signs, but the analysis of the critical points of (2) is rather delicate, because the location variables and all the relative scales remain coupled. This is one of the reasons why explicit constructions in natural geometries have been obtained only for particular configurations. For example, in [16] a family with an even number of bubbles placed at the vertices of a regular polygon with alternating signs is constructed in the unit ball: the symmetry reduces all the scales to one parameter and all the locations to one radial variable. The same paper also produces many sign-changing patterns in suitable dumbbell-shaped domains, and treats the case of one positive and one negative bubble in an arbitrary smooth bounded domain. In all these cases the difficulty is not the reduction itself, but the analysis of the critical points of the finite-dimensional function. In dimension N=4N=4 the situation is different: the concentration scales are exponentially small in 1/ε1/\varepsilon and the leading equations for the relative scales become spectral. Hence the scale variables can be eliminated through the spectral structure of the interaction matrix, which makes accessible configurations that are considerably harder to detect through the higher-dimensional reduced function.

This spectral mechanism was exploited in [18], where several families of sign-changing multi-peak solutions were constructed in dimension four, in configurations whose sign pattern is not alternating and which are therefore not covered by the symmetric constructions of [16]. Among them: in the unit ball, a positive bubble at the centre surrounded by kk negative bubbles placed at the vertices of a regular polygon, and positive and negative bubbles placed at the vertices of two regular polygons lying in orthogonal planes; in an annulus, bubbles with alternating signs at the vertices of a regular polygon. Whether these configurations survive in higher dimensions was left open in [18]. When N≥5N \ge5 the scales are only polynomially small in ε\varepsilon, the self-interaction term bN∑jdj2b_N\sum_j d_j^2 in (2) is of the same order as the interaction terms, and the scales can no longer be eliminated by a spectral argument: the location of the bubbles and their relative scales must be determined together, as critical points of the full function (2).

The aim of this paper is to show the persistence of the multi-peak configurations constructed in [18] for N=4N=4 in dimensions N≥5N\ge5, where no spectral mechanism is available and one has to deal with the full function (2). We focus on a particularly interesting one: in the unit ball we look for solutions with one positive peak at the centre, surrounded by an arbitrary number k≥2k\ge2 of negative peaks placed at the vertices of a regular polygon.

To state our results we introduce the relevant symmetries. Let Γk⊂O(N)\Gamma_k\subset O(N) be the group generated by the rotation of angle 2π/k2\pi/k in the (x1,x2)(x_1,x_2)-plane, by the reflection (x1,x2,x′′)↦(x1,−x2,x′′)(x_1,x_2,x'')\mapsto(x_1,-x_2,x'') and by O(N−2)O(N-2) acting on x′′=(x3,…,xN)x''=(x_3,\ldots,x_N), and let

Hk:={u∈H01(B1):u∘γ=u∀γ∈Γk},\mathcal{H}_k:=\{u\in H_0^1(\mathcal{B}_1):u\circ\gamma=u\quad\forall\gamma\in\Gamma_k\},

where B1\mathcal{B}_1 denotes the unit ball of RN\mathbb{R}^N centred at the origin. For r∈(0,1)r\in(0,1) we consider the vertices of a regular kk-gon of radius rr in the (x1,x2)(x_1,x_2)-plane,

ξj=ξj(r):=r(cos⁡2π(j−1)k,sin⁡2π(j−1)k,0,…,0),j=1,…,k,(4)\xi_j=\xi_j(r):=r\left(\cos\frac{2\pi(j-1)}{k},\sin\frac{2\pi(j-1)}{k},0,\ldots,0\right),\qquad j=1,\ldots,k, \tag*{(4)}

and we set M:=(0,1)×(0,∞)2\mathcal{M} := (0,1) \times(0,\infty)^2. We look for solutions of the form

References

  1. [1]A. L. Amadori, F. Gliadiali, M. Grossi, A. Pistoia, G. Vaira, A complete scenario on nodal radial solutions to the Brezis–Nirenberg problem in low dimensions, Nonlinearity 34 (2021), no. 11, 8055–8093.
  2. [2]T. Aubin, Espaces de Sobolev sur les variétés riemanniennes, Bull. Sci. Math. (2) 100 (1976), 149–173.
  3. [3]T. Bartsch, T. D’Aprile, A. Pistoia, On the profile of sign changing solutions of an almost critical problem in the ball, Bull. Lond. Math. Soc. 45 (2013), no. 6, 1246–1258.
  4. [4]T. Bartsch, A. M. Micheletti, A. Pistoia, On the existence and the profile of nodal solutions of elliptic equations involving critical growth, Calc. Var. Partial Differential Equations 26 (2006), no. 3, 265–282.DOI
  5. [5]H. Brezis, L. Nirenberg, Positive solutions of nonlinear elliptic equations involving critical Sobolev exponents, Comm. Pure Appl. Math. 36 (1983), no. 4, 437–477.DOI
  6. [6]L. A. Caffarelli, B. Gidas, J. Spruck, Asymptotic symmetry and local behavior of semilinear elliptic equations with critical Sobolev growth, Comm. Pure Appl. Math. 42 (1989), no. 3, 271–297.DOI
  7. [7]O. Druet, Elliptic equations with critical Sobolev exponents in dimension 3, Ann. Inst. H. Poincaré C Anal. Non Linéaire 19 (2002), no. 2, 125–142.
  8. [8]P. Esposito, N. Ghoussoub, A. Pistoia, G. Vaira, Sign-changing solutions for critical equations with Hardy potential, Anal. PDE 14 (2021), no. 2, 533–566.DOI
  9. [9]Z.-C. Han, Asymptotic approach to singular solutions for nonlinear elliptic equations involving critical Sobolev exponent, Ann. Inst. H. Poincaré C Anal. Non Linéaire 8 (1991), no. 2, 159–174.DOI
  10. [10]A. Iacopetti, G. Vaira, Sign-changing tower of bubbles for the Brezis–Nirenberg problem, Commun. Contemp. Math. 18 (2016), no. 1, 1550036, 53 pp.arxiv.org/abs/1402.1451
  11. [11]A. Iacopetti, G. Vaira, Sign-changing blowing-up solutions for the Brezis–Nirenberg problem in dimension four and five, Ann. Sc. Norm. Super. Pisa Cl. Sci. 18 (2018), 1–38.
  12. [12]T. König, P. Laurain, Fine multibubble analysis in the higher-dimensional Brezis–Nirenberg problem, Ann. Inst. H. Poincaré C Anal. Non Linéaire 41 (2024), no. 5, 1239–1287.
  13. [13]T. König, P. Laurain, Multibubble blow-up analysis for the Brezis–Nirenberg problem in three dimensions, arXiv:2208.12337 (2022).arxiv.org/abs/2208.12337
  14. [14]F. Liu, G. Vaira, J. Wei, Y. Wu, Construction of bubbling solutions of the Brezis–Nirenberg problem in general bounded domains (I): the dimensions 4 and 5, J. Lond. Math. Soc. 112 (2025), no. 2, e70246.
  15. [15]F. Liu, G. Vaira, J. Wei, Y. Wu, On the Brezis–Nirenberg problems: open questions and new results in dimension six, Discrete Contin. Dyn. Syst. 54 (2026), 151–176.
  16. [16]A. M. Micheletti, A. Pistoia, On the effect of the domain geometry on the existence of sign changing solutions to elliptic problems with critical and supercritical growth, Nonlinearity 17 (2004), 851–866.DOI
  17. [17]M. Musso, A. Pistoia, Multispike solutions for a nonlinear elliptic problem involving the critical Sobolev exponent, Indiana Univ. Math. J. 51 (2002), no. 3, 541–579.
  18. [18]A. Pistoia, G. M. Rago, G. Vaira, Sign-changing multi-bubble solutions for the Brezis–Nirenberg problem in four dimensions, arXiv:2608.21239.
  19. [19]A. Pistoia, G. Vaira, Nodal solutions of the Brezis–Nirenberg problem in dimension 6, Anal. Theory Appl. 38 (2022), no. 1, 1–25.arxiv.org/abs/2010.09026
  20. [20]B. Premoselli, Towers of bubbles for Yamabe-type equations and for the Brézis–Nirenberg problem in dimensions n ≥ 7, J. Geom. Anal. 32 (2022), no. 3, Paper No. 73, 65 pp.
  21. [21]G. M. Rago, Sign-changing bubbling solutions on polygons for the Brezis–Nirenberg problem in high dimensions, preprint.
  22. [22]O. Rey, The role of the Green’s function in a nonlinear elliptic equation involving the critical Sobolev exponent, J. Funct. Anal. 89 (1990), no. 1, 1–52.DOI
  23. [23]G. Talenti, Best constant in Sobolev inequality, Ann. Mat. Pura Appl. (4) 110 (1976), 353–372.DOI

Paper details

vlnc:W_368nmjcb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

ChatGPT (Open AI)

View statement
Looking for related papers…
Contents