CUBO A Mathematical Journal Vol.15, No¯ 01, (151–158). March 2013 On the Poisson’s equation −∆u = ∞. Carlos Cesar Aranda 1 Blue Angel Navire Research Laboratory, Rue Eddy 113 Gatineau QC Canada. carloscesar.aranda@gmail.com ABSTRACT Let Ω ⊂ RN be a bounded domain. We proof the existence of a bounded solution of the Poisson’s equation −∆u = ∞ on Ω. RESUMEN Sea Ω ⊂ RN un dominio acotado. Probamos la existencia de una solución acotada para la ecuación de Poisson −∆u = ∞ en Ω. Keywords and Phrases: Newtonian potential; nonlinear analysis; celestial mechanics 2010 AMS Mathematics Subject Classification: 35J25, 35J60. 1Dedicated to Professor Gaston M N’Guérékata on the occasion of his 60th birthday. 152 Carlos C. Aranda CUBO 15, 1 (2013) 1 Introduction. In [19] it is stated that Le mouvement d’un corps libre consiste dans le mouvement de translation de son centre de gravite et dans le changement de sa position autour de ce point. La recherche du mouvument du centre de gravité se réduit à déterminer le mouvement d’un point sollicité par des forces donnés; et, relativament aux corps célestes, ces forces sont le résultat des attractions de spheroides dont la figure est supposée connu. Soient dm une molécule d’un sphéroide; x′, y′, z′ les trois coordennées orthogonales de cette molécule; dm sera de la forme ξdx′dy′dz′, ξ étant fontion de x′, y′, z′. Soient encore x, y, z les coordonnées d’un point attir’e, on aura V = ∫ ξdx′dy′dz′ √ (x′ − x)2 + (y′ − y)2 + (x′ − y)2 (1) cette intégrale étant prise relativementà toute l’étendue du sphéroide. Ses limites étant indépendantes de x, y, z ainsi que les variables x′, y′, z′, il est clair qu’en differential l’expression de V par rapport àx, y, z il suffira, dans cette différentiation, d’avoir égard au radical que renferme cette expression, et alors il est facile de voir que l’on a 0 = ∂2V ∂x2 + ∂2V ∂y2 + ∂2V ∂y2 . (2) In modern interpretation of potential V of mass distributions, we have V(x, y, z) = ∫ G ξ(x′, y′, z′)dx′dy′dz′ √ (x′ − x)2 + (y′ − y)2 + (x′ − y)2 . (3) where ξ(x′, y′, z′) is the density of a mass distribution in the space x′, y′, z′. Then ∇V furnishes the gravity field force and −∆V = 0 on R3 − G. In 1813 Poisson found that for a ball G the following equation is valid in the case of constant density ξ(x, y, z′) = ρ −∆u = 4πρ on G Poisson’s equation. Therefore a natural question is: there exists a solution for Poisson’s equation with ρ = ∞?. That kind of solution will be related to gravity potential of bodies with infinite density or black holes. The authors are not aware of a previous result deducing the existence of black holes using Newton gravity theory or the gravity potential inside of a black hole. The equation − ∆u = up, (4) for p a nonnegative real number and u > 0in a Ball of radius R in R3, with Dirichlet boundary conditions was introduced by Lane [18] for modelling both the temperature and the density of CUBO 15, 1 (2013) On the Poisson’s equation −∆u = ∞. 153 mass on the surface of the sun. Today the problem (4) is named Lane-Emden-Fowler equation. It was used first in the mid-19th century in the study of internal structure of stars mainly by Chandrasekhar [4, 7, 9]. Singular Lane-Emden-Fowler equations (p < 0) has been considered in a remarkable pioneering paper by Fulks and Maybe [10]. Eddigton [6] proposed the equation − ∆u = exp(2u) 1+ | x |2 in R3, (5) in order to represent the gravitational potential u of a globular cluster of stars. Matukuma [20] introduced the equation − ∆u = ur 1+ | x |2 in R3, (6) where u is the gravity potential, ρ = (2π)−1(1+ | x |2)−1ur is the density and ∫ R3 ρdx is the total mass to study the gravitational potential u of a globular cluster of stars. For the same problem Hénon [15] suggested − ∆u =| x |l ur in Ω ⊂ R3. (7) Black holes solutions means that the gravitational potential of the cluster behaves like 1 r (r =| x |) near the center. Peebles [16, 17] gives for the first time a derivation of the steady state distribution of the star near a massive collapsed object. The question of the existence of black hole in a globular cluster is still open (1995). Core collapse does occur, for instance using Hubble Space Telescope, Bendinelli et.al. [2] reported the first detection of a collapsed core globular cluster in M31. On May 25, 1994 astronomers at NASA headquarters announced the Hubble Space Telescope finding of a supermassive black hole in the heart of the giant galaxy M87, more than 50 million light-years. The equation − ∆ 1 | x − x0 | = 4πδ(x − x0) in R 3, has a deep insight because relate the formulation of the Laplace operator and the Dirac δ function in a weak sense. The Laplace operator with point interaction in R3 given by −∆ + αδ, α ∈ R has been widely study for your applications in quantum physics (see for expample [11]) and in seismic imaging [3]. Our purpose in this paper is to give a classical interpretation to the equation − ∆u = ∞ in Ω ⊂ RN. (8) We define: 154 Carlos C. Aranda CUBO 15, 1 (2013) Definition 1.1. The equation (8) has a classical solution if there exist two non decreasing sequences of functions {uj} ∞ j=1 ∈ C(Ω) ∩ C 2(Ω) and {fj} ∞ j=1 such that −∆uj = fj in Ω, and limj→∞ fj(x) = ∞ for all x ∈ Ω and limj→∞ uj(x) = u(x) < ∞ for all x ∈ Ω. Our main result in this article is as follows. Theorem 1.2. Let Ω be a bounded domain in RN, N ≥ 3. Then the problem −∆u = ∞ in Ω, (9) has a non negative classical solution u. Under the authors knowledge this is the first compactness result dealing with infinite on a non trivial domain (see for example [21] first chapter: direct methods in the calculus of variations). Similarly the theory of generalized functions not allow solutions to this kind of problem because every distribution is locally a Newtonian potential: Theorem 1.3. (page 277 [5]) Let Ω be an open set of RN, f ∈ D′(Ω) and u a solution (in the sense of distributions) of Poisson’s equation ∆u = f on Ω . Then for every bounded open set Ω1 with Ω1 ⊂ Ω there exists f1 ∈ E ′ the space of distributions on RN with compact support, such that f1 = f on Ω and u = the Newtonian potential of f1 on Ω1. Moreover if we study this problem using a weak formulation in Sobolev’s spaces, the Georgi- Nash-Moser theory cannot be used to derive any comparable compacity result [14]. We will use a non linear singular elliptic approach as in [1, 8, 13] to obtain the result. Our strategy is study the auxiliary problem −∆uǫ,m = gm(uǫ) in Ω, uǫ,m = ǫ on ∂Ω, where gm : (0, ∞) → (0, ∞), m = 1, . . . ∞ is non increasing locally Hölder continuous function singular at the origin with the properties gm(s) = g(s) for all s ≥ 1 and limm→∞ gm(s) = ∞ for all s ∈ (0, 1), m = 1, . . . , ∞ and g : (0, ∞) → (0, ∞) is strictly non increasing locally Hölder continuous function singular at the origin. Our result 1.2 is obtained letting limm→∞,ǫ→0+ uǫ,m. This limit by definition has not weak derivatives of first or second order. CUBO 15, 1 (2013) On the Poisson’s equation −∆u = ∞. 155 2 Auxiliary results Theorem 2.1 ([1]). Let Ω be a smooth bounded domain in RN, N ≥ 3, g : (0, ∞) → (0, ∞) is non increasing locally Hölder continuous function (that may be singular at the origin). Then the problem −∆uǫ = g(uǫ) in Ω, uǫ = ǫ on ∂Ω, has a unique positive solution u ∈ C(Ω) ∩ C2(Ω) for ǫ ≥ 0. Moreover uǫ2 ≥ uǫ1 for ǫ2 ≥ ǫ1. We consider the the auxiliary problem −∆um = gm(um) in Ω, um = 0 on ∂Ω, (10) Lemma 2.2. Let um be a solution of the equation (10). Then um+j ≥ um. Proof. Suppose that there exists x0 ∈ Ω such that um(x0) > um+j(x0). Therefore for τ > 0 small enough we have the inequality um(x0) > τ + um+j(x0). Then by continuity in Ω of the function F(x) = um(x)−τ−um+j(x) there exist a non empty open set Ωτ such that F(x) > 0 for all x ∈ Ωτ and F = 0 on ∂Ωτ. Using that um(x) > τ + um+j(x) for all x ∈ Ωτ, we deduce gm(um(x)) ≤ gm+j(um(x)) ≤ gm+j(τ + um+j(x)) ≤ gm+j(um+j(x)) for all x ∈ Ωτ. Then −∆um ≤ −∆(um+j + τ) in Ωτ, um = um+j + τ on ∂Ωτ. and we obtain um ≤ um+j + τ in Ωτ (Theorem 3.3 [14]) a contradiction. Lemma 2.3. Let um be a solution of the equation (10). Then gm+j(um+j(x)) ≥ gm(um(x)). Proof. Suppose that there exists x0 ∈ Ω such that gm(um(x0)) > gm+j(um+j(x0)). Then by continuity in Ω of the function H(x) = gm(um(x)) − gm+j(um+j(x)), there exists Ω̂ ⊂ Ω such that H(x) > 0 in Ω̂ and H(x) = 0 on ∂Ω̂ − ∆um ≥ −∆um+j in Ω̂, um = um+j on ∂Ω̂. We imply um ≥ um+j in Ω̂ (Theorem 3.3 [14]). Therefore gm(um(x)) ≤ gm(um+j(x)) ≤ gm+j(um+j(x)) for all x ∈ Ω̂. A contradiction. Remark 2.4. In the proof of Lemmas 2.2 and 2.3 it is assumed only that gm is a non increasing continuous function. 156 Carlos C. Aranda CUBO 15, 1 (2013) 3 Proof Proof of Theorem 1.2. Let us consider the problem −∆v = g(v) in Ω, v = 0 on ∂Ω. We introduce the equations −∆e = g(e) in Ω, e = 1 on ∂Ω. −∆w = g(e) in Ω, w = 0 on ∂Ω. Using v ≤ e (see Lemma 2.3 and 2.6 in [1]), we infer −∆w = g(e) ≤ g(v) = −∆v in Ω, w = 0 = v on ∂Ω. Then w ≤ v in Ω . Setting g0 = g and using the auxiliary results with the new sequence {gj} ∞ j=0, we conclude that w ≤ um ≤ e for m = 1, . . . ∞. Using Lemma 2.2, we infer the existence of limm→∞ um(x) = u(x) for all x ∈ Ω. We restrict ourselves to the situation Ω = B1(0) where B1(0) is the ball of radius 1 with center at 0. Applying the main result of [12] we infer that um is a radial function with ∂um ∂r < 0. Therefore u is also a radial non increasing function. We proceed by contradiction, suppose that lim m→∞ gm(um(x)) < ∞ for all 0 ≤‖ x ‖< 1. Our first implication is that the function u is strictly non increasing, because if exists (r1, r2) with r2 < 1, and u(r1) = u(r2). Then −∆u = 0 on the annulus A(r1, r2). Using Theorem 9.11 page 235 in [14], we deduce ‖ um ‖H2,p(Ω′)≤ C(N, p, Ω ′, A(r1, r2))(‖ um ‖Lp(A(r1,r2)) + ‖ g(um) ‖Lp(A(r1,r2))) ≤ C(N, p, Ω′, A(r1, r2))(‖ e ‖Lp(A(r1,r2)) + ‖ lim sup m→∞ gm(um(r2)) ‖Lp(A(r1,r2))), for all p > N, therefore by Sobolev’s embedding theorem (Theo. 7.26 [14]) we deduce ‖ um ‖C1,α(Ω′)≤ C. We use a non negative test function ϕ with support contained in Ω′: 0 = ∫ Ω′ ∇u · ∇ϕdx = lim m→∞ ∫ Ω′ ∇um · ∇ϕdx = ∫ Ω′ gm(um)ϕdx ≥ ∫ Ω′ g0(u0)ϕdx > 0. CUBO 15, 1 (2013) On the Poisson’s equation −∆u = ∞. 157 Contradiction, therefore we deduce that u is a strictly non increasing function. Moreover using again estimates in Theorems 9.11 and 9.12 in [14] we have u ∈ C1,αloc(B1(0)). By assumption lim supr→1 u(r) ≥ 1, therefore u(r) > 1 for 0 ≤ r < 1. By construction there exists 0 < r0 < 1 such that g0(u0(r0)) > g0(1). Using Lemma 2.3, we derive g0(u0(r0)) ≤ gm(um(r0)). But limm→∞ um(r0) = u(r0) > 1 and therefore for m big enough um(r0) > 1. Moreover gm(um(r0)) = g0(um(r0)) < g0(1) because g0 is strictly non increasing. Contradiction. It is follows that there exists 0 ≤ r1 < 1 such that lim m→∞ gm(um(r1)) = ∞. Now, because um is a radial non increasing function, we infer that gm(um(r1)) ≤ gm(um(r)) for all r1 < r < 1. So lim m→∞ gm(um(r)) = ∞ for all r1 ≤ r < 1. Now for Ω a bounded domain in RN, N ≥ 3 consider the transformation um( a+x R ). This end the proof. Received: October 2012. Revised: February 2013. References [1] C. C. Aranda and T. Godoy. Existence and Multiplicity of positive solutions for a singular problem associated to the p-laplacian operator. Electron. J. Diff. Eqns., Vol. 2004(2004), No. 132, pp. 1-15 [2] O. Bendinelli, C. Cacciari, S. Djorgorski, L. Federici, ER. Ferraro, Fi Fusi Pecci, G. Parmeg- giani, N. Weir and E Zaratti, “The first detection of a collapsed core globular cluster in M31”. Astrophy. I., 409 (1993), LI7-LI9. [3] N. Bleistein, Mathematics of modelling, migration and inversion with Gaussian beams. Center for wave phenomena, Colorado School of Mines, 2009. [4] S. Chandrasekhar, An introduction to the study of stellar structure. Dover, New York, 1957. [5] R. Dautray and J. L. Lions Mathematical analysis and numerical methods for science and technology. Vol. 1 Physical Origins and classical methods. Springer Verlag 1990. [6] A.S. Eddington, The dynamics of a globular stellar system. Monthly Notices Roy. Ast. Soc., 75 (1915), 366-376. 158 Carlos C. Aranda CUBO 15, 1 (2013) [7] V. R. Emden. Gaskugeln. Teubner, Leipzig, 1907. [8] J. Hérnandez and F. J. Mancebo. Singular elliptic and parabolic equations. In Handbook of differential equations (ed. M. Chipot and P. Quittner), vol 3 (Elsevier 2006) [9] R. H. Fowler. Further studies of Emden’s and similar differential equations. Q. J. Math. (Ox- ford Series) 2 (1931), 259-288. [10] W. Fulks and J. S. Maybe. A singular nonlinear equation. Osaka J. Math. 12 (1960), 1-19. [11] P. Kurasov Singular and superlinear perturbation Hilbert space methods. Contemporary Math- ematics Vol. 340, 2004. [12] B. Gidas, Wei Ming Ni and L. Nirenberg. Symmetry and related properties via the maximum principle. Comm. Math. Phys. Volume 68 number 3(1979), 209-243. [13] M. Ghergu and V. Rădulescu. Singular Elliptic Problems: Bifurcation & Asymptotic Analysis. Oxford Lecture Series in Mathematics and Its Applications, 2008. [14] David Gilbarg and Neil S. Trudinger Elliptic partial differential equations of second order. Classics in mathematics reprint of 1998 edition Springer [15] M. Hénon. Numerical experiments on the stability of spherical stellar systems. Astronomy and Astrophysics 24 (1973), 229-238. [16] P.I.E. Peebles,Black holes are where you find them. Gen. Rel. Grav., 3 (1972), 63-82. [17] P.I.E. Peebles,Star distribution near a collapsed object. Ast. I., 178 (1972), 371-375. [18] J. Homer Lane. On the Theoretical Temperature of the Sun under the Hypothesis of a gaseous mass Maintining Its Volume by Its Internal Heat and Depending on the Law of Gases Known to Terrestrial Experiment. Amer. J. Sci, 2d ser. 50 (1869), 57-74. [19] P. S. de Laplace Memoire sur les mouvements des corps celestes autout de leurs centre de gravite, Ouvres completes tome douzieme. [20] T. Matukuma. Sur la dynamique des amas globulaires stellaires. Proc. Imp. Acad. 6 (1930), 133-136. [21] M. Struwe. Variational methods: applications to nonlinear partial differential equations and Hamiltonian systems. Springer 2000.