Question on Theorem 2.2.13 in Partial Differential Equations by Evans

52 Views Asked by At

I'm working through the section on Laplace's Equation in Partial Differential Equations by Evans. I'm having trouble following a step in Evans's proof of the symmetry of Green's function (Theorem 2.2.13 on page 35). The statement of the theorem is as follows: For all $x,y \space \epsilon \space U$ with $x \not = y$ we have $$ G(y,x)=G(x,y)$$ The function $G$ is Green's function which is defined as $G(x,y) = \Phi(y-x)-\phi^x(y) $ where $\Phi $ is the fundamental solution of Laplace's equation and $\phi^x(y)$ is constructed such that $\Delta \phi^x = 0 $ in $U$ and $\phi^x = \Phi(y-x)$ on $\partial U $. He begins the proof by defining $v(z) := G(x,z)$ for a fixed $x$ and $ w(z) := G(y,z) $ for a fixed $y$.

In the middle of the proof, he states that $v(z) = \Phi(z-x) - \phi^x(z)$, where $\phi^x$ is smooth in $U$ and that therefore $$(*) \space \space\lim_{\varepsilon \to 0}\int_{\partial B(x,\varepsilon)}\frac{\partial v}{\partial \nu}wdS = \lim_{\varepsilon \to 0}\int_{\partial B(x,\varepsilon)}\frac{\partial \Phi}{\partial \nu}(x-z)w(z)dS = w(x)$$ Here, $\frac{\partial v}{\partial \nu} = \nabla v \cdot \nu $ where $\nu$ is the outward oriented unit normal vector. It is here that I get confused. I understand the second equality in (*) but I don't understand how the first equality works. By my understanding, $$\frac{\partial v}{\partial \nu} = \nabla v \cdot \nu = \nabla \Phi(y-x) \cdot \nu - \nabla \phi^x \cdot \nu$$ and I don't understand what happens to the $\nabla \phi^x$. Why isn't it this instead?: $$\lim_{\varepsilon \to 0}\int_{\partial B(x,\varepsilon)}\frac{\partial v}{\partial \nu}wdS = \lim_{\varepsilon \to 0}\int_{\partial B(x,\varepsilon)}[\frac{\partial \Phi}{\partial \nu}(x-z)-\frac{\partial \phi^x}{\partial \nu}]w(z)dS$$

Any help would be appreciated.