I would like you to follow my logic, confirming it if correct, suggesting change when flawed or suboptimal.

Let $p$ be a complex polynomial \begin{gather*} p:\mathbb{C}\longrightarrow\mathbb{C},\\ \deg p = n,\quad n\in\mathbb{N}. \end{gather*} Let $z=x+iy$, and write $p(z)=u(x,y)+iv(x,y)$. Note that $u$ and $v$ are harmonic functions. Define the set $\mathcal{R}=\{z\in\mathbb{C}:|p(z)|\leq R\}$, where $R>0$ is chosen sufficiently large.

  • $\partial\mathcal{R}=\{z\in\mathbb{C}:|p(z)|=R\}$ by continuity and the maximum modulus principle.

  • Hence $\mathcal{R}$ is closed. Since $|p(z)|\rightarrow\infty$ as $|z|\rightarrow\infty$, the set $\mathcal{R}$ is also bounded.

  • $\mathbb{C}\setminus\mathcal{R}$ is open and by the maximum modulus principle necessarily unbounded.

  • $\operatorname{int}\mathcal{R}=\mathcal{R}\setminus\partial\mathcal{R}$ is open and bounded. It's also simply connected by the maximum modulus principle.

  • $\mathcal{R}_{1/2}=\left\lbrace z\in\mathbb{C}: |p(z)|\leq\frac{R}{2}\right\rbrace\subset\operatorname{int}\mathcal{R}$ by continuity of $|p(z)|$.

  • Hence $\partial\mathcal{R}\subset\mathbb{C}\setminus\mathcal{R}_{1/2}=\left\lbrace k\in\mathbb{C}: |p(k)|>\frac{R}{2}\right\rbrace$.

Now suppose $R$ is so large that \begin{align*} p^{\prime}(z)\neq0\text{ when }z\in\mathbb{C}\setminus\mathcal{R}_{1/2}. \end{align*}

  • Then the boundary $\partial\mathcal{R}$ is given by a smooth curve: Suppose $x_{0}+iy_{0}=z_{0}\in\partial\mathcal{R}$. The condition on the boundary is $|p(z_{0})|=R$ and $p^{\prime}(z_{0})\neq0$. Let $F(x,y)=(u(x,y))^{2}+(v(x,y))^{2}-R^{2}$. The zero level set of $F$ is precisely the set $\partial\mathcal{R}$. Note that \begin{align*} \nabla F = 2\, \underbrace{\begin{pmatrix} \partial_{x}u & \partial_{x}v \\ \partial_{y}u & \partial_{y}v \end{pmatrix}}_{A}\, \underbrace{\begin{pmatrix} u\\ v \end{pmatrix}}_{\bar{u}}, \end{align*} where $\det A=p^{\prime}\neq0$ at the point $z_{0}$. Since the matrix $A$ is invertible at the point $z_{0}$, the only element in it's nullspace is the zero vector. However $\left\vert\bar{u}\right\vert^{2}=R^{2}$ at $z_{0}$. That meas $\left\vert\nabla F\right\vert\neq0$ at $z_{0}$, hence atleast one of the partial derivatives $\partial_{x}F$ or $\partial_{y}F$ is non zero at the point $z_{0}$. So the implicit function theorem gives that there exists a neighbourhood of the point $z_{0}$ such that the set of $z$ such that $F(z)=0$ in that neighbourhood consists of a smooth curve passing through $z_{0}$.
  • The above smooth curve is bounded.

Here are some things I'm wondering about

  • Is the smooth curve giving the boundary $\partial\mathcal{R}$ simple (for big enough $R$)? I also think it's closed but not sure how to show it.

  • I believe there is some upper and lower bound on $\partial\mathcal{R}$ for large $R$ in the form of disks centred at the origin. That is \begin{align*} \mathcal{R}_{small}&=\{z\in\mathbb{C}:|z|<R_{small}\},\\ \mathcal{R}_{big}&=\{z\in\mathbb{C}:|z|<R_{big}\}, \end{align*} where $R_{big}>R_{small}>0$, such that $\mathcal{R}_{small}\cap\partial\mathcal{R}=\emptyset$ and $\mathcal{R}_{big}\cap\partial\mathcal{R}=\partial\mathcal{R}$. What is the biggest $\mathcal{R}_{small}$ and smallest $\mathcal{R}_{big}$?

  • Are there some more interesting properties to discuss?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.