1
$\begingroup$

Let $f: \mathbb{R}^{3} \to \mathbb{R}$ be the solution of the following PDE: $$\Delta f -\frac{1}{2}h f = 0$$ where $h \in C_{c}^{\infty}(\mathbb{R}^{3})$ (compactly supported an smooth) and $f$ satisfies $\lim_{|x|\to \infty}f(x) = 1$.

I am trying to prove the following: $f$ behaves asymptotically as: $$f(x) = 1 -\frac{C}{|x|}$$ for some constant $C > 0$ and this $C$ is exactly the solution of: $$8 \pi C = \int_{\mathbb{R}^{d}}h(x)f(x).$$

I honestly don't know where to begin.

$\endgroup$
6
  • 1
    $\begingroup$ Why do you think it is true? $\endgroup$ Commented Sep 19, 2023 at 21:06
  • 4
    $\begingroup$ It should follow by writing $f=1+v$, $\Delta v=h(1+v)$ and expressing $v$ as the Newtonian potential of $h(1+v)=hf$. $\endgroup$ Commented Sep 19, 2023 at 21:17
  • $\begingroup$ @GiorgioMetafune an equation like this one usually arises in quantum mechanics. $\endgroup$ Commented Sep 19, 2023 at 21:50
  • $\begingroup$ That type of questions arise in general relativity too. You should have a look at the paper by R. Bartnik "The mass of an asymptotically flat manifold" $\endgroup$ Commented Sep 20, 2023 at 13:02
  • 1
    $\begingroup$ @RomainGicquaud: that's swatting a fly with a bazooka. $\endgroup$ Commented Sep 20, 2023 at 13:08

2 Answers 2

3
$\begingroup$

For the specific question asked, just follow what Giorgio said.

Let $\phi = f-1$, then $\Delta \phi = \frac12 hf$ is a compactly supported function on $\mathbb{R}^3$.

Now let $g$ be the Newton potential of $\frac12 hf$ (so $g = c(\frac12 hf)*\frac{1}{|x|}$ for some appropriate constant $c$), then you also have $\Delta g = \frac12 hf$.

$g$ is well-defined (as $\frac12hf$ has compact support and is continuous), and you can check that it decays to $0$ at infinity.

Thus $\phi - g$ is a harmonic function that vanishes at infinity, and hence by the maximum principle is zero. This shows that $f-1 = g$.

The asymptotics comes straight out of the convolution definition for $g$.

$\endgroup$
4
  • $\begingroup$ @ChristianRemling: I interpret "behave asymptotically" means $f(x) = 1 + C/|x| + o(1/|x|)$ as $|x|\to\infty$. Your interpretation seems too rigid: even if there were a solution that is exactly $1 + C/|x|$ outside a compact set, just translating the whole thing sideways will give immediately a counterexample. $\endgroup$ Commented Sep 21, 2023 at 1:37
  • $\begingroup$ @ChristianRemling: also, I read it neither as about existence nor uniqueness, but about a priori estimate. (Maybe the OP can clarify...) $\endgroup$ Commented Sep 21, 2023 at 1:38
  • $\begingroup$ Thanks for your answer Willie. Indeed, by behave asymptotically I meant $f(x) = C/|x|+o(1/|x|)$. I still wanted to clarify the second expression for $8\pi C$. It is enough to take a large enough closed ball, say, $B_{r}(0)$, and use the maximum principle for harmonic functions? The factor $8\pi$ probably comes by integrating the $\theta, \varphi$ angles in spherical coordinates. $\endgroup$ Commented Sep 21, 2023 at 7:14
  • $\begingroup$ @IamWill: $8\pi = 4\pi \times 2$, the $2$ comes from the $\frac12$ in your equation. The $4\pi$ is the surface area of the unit 2 sphere, and is the constant in Newton potential. $\endgroup$ Commented Sep 21, 2023 at 13:47
2
$\begingroup$

I think you can get the result blowing-down the solution. I'll just sketch the argument since I don't want to be confusing with the details.

Set $u:=1-f$, then $u$ solves $-\Delta u = \frac{h}{2}(1-u)$ and satisfies $\lim_{|x|\to \infty} u(x)=0$. For $r>0$ consider $u_r(x):= ru(rx)$, you can check easily that $u_r$ solves $$-\Delta u_r =\frac{r^3}{2}h(rx)(1-u_r) =: g_r(x)\,. $$ Since $h$ has compact support you have $-\Delta u_r= g_r \to 0$ in $C^2_{\rm loc}(\mathbb{R}^3 \setminus \{0\})$ and moreover $$ \int_{\mathbb{R}^3} g_r \, dx = \frac{r^3}{2} \int_{\mathbb{R}^3} h(rx)(1-u_r(x)) \, dx = \frac{1}{2}\int_{\mathbb{R}^3} h(1-u) = \frac{1}{2}\int_{\mathbb{R}^3} hf =: c_0 \,.$$ Hence $g_r \to c_0\, \delta_{\{x=0\}}$ weakly in the sense of distributions, and from here by elliptic estimates you get that $u_r \to c_0 \, G_3$ where $G_3=C_n/|x|$ is the fundamental solution in $\mathbb{R}^3$. This holds since $G_3$ is the unique solution to $-\Delta G_3=\delta_{\{x=0\}}$ that goes to zero as $|x|\to \infty$. This implies that, for example, in the annulus $A_{1,2}:=B_2(0)\setminus B_1(0)$ there holds $$ \sup_{x\in A_{1,2}} \left|ru(rx) -\frac{c_0 C_n}{|x|} \right| \to 0$$ as $r\to \infty$, or equivalently $$ r \sup_{x \in A_{r,2r}} \left|u(x) -\frac{c_0 C_n}{|x|} \right| \to 0 \,.$$ This implies what you ask.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .